Stable Isotope Geochemistry
 0939950553

Citation preview

— STABLE ISOTPOE GEOCHEMISTRY — 43

Reviews in Mineralogy and Geochemistry

43

FOREWORD The scientific editors of this volume, John Valley and David Cole, organized a Short Course on Stable Isotope Geochemistry presented November 2-4 2001 in conjunction with the annual meetings of the Geological Society of America in Boston, Massachusetts. The contributors to this review volume were the instructors. The Mineralogical Society of America (MSA) sponsored the course and is published and distributor of this and other books in the Review in Mineralogy and Geochemistry series. Alex Speer, Executive Director of MSA, and Myrna Byer, of Printing Service Associates, Inc., are acknowledged for their considerable contributions to the success of this volume. MSA has been in partnership with the Geochemical Society (GS) since 2000 and the publication of Volume 39, Transformation Processes in Minerals. As a result, the Reviews volumes are now covering an ever-widening range of subjects. An Additional series editor, Jodi J Rosso, has joined the team to manage those short-course publications and other books which will be sponsored by the Geochemical Society MSA and the editors of Stable Isotope Geochemistry are particularly grateful to the Geosciences Research Program (Nick Woodwood, director) of the U.S. Department of Energy for financial support of this publications and student scholarships through a grant to MSA. Paul H. Ribbe Series Editor for MSA Blacksburg, Virginia September 14, 2001 PREFACE This volume follows the 1986 reviews in Mineralogy (Vol.16) in approach but reflects significant changes in the fields of Stable Isotope Geochemistry , In terms of new technology, new sub-disciplines, and numbers of researchers, the field has changes more in the past decade than in any other since that of it birth. Unlike the 1986 volume, which was restricted to high temperature fields, this book covers a wider range of disciplines. However, it would not be possible to fit a comprehensive review into a single volume. Our goal is to provide state-ofthe-art reviews in chosen subjects that have emerged or advanced greatly since 1986. The field of Stable Isotope Geochemistry was born of a good idea and nurtured by technology, in 1947, Harold Urey published his calculated values of reduced partition function for oxygen isotopes and his idea (a good one!) that the fractionation of oxygen isotope between calcite and water might provide a means to estimate the temperatures of geologic events. Building on wartime Advances in electronics, Alfred Nier then designed and built the dual-inlet, gas source mass-spectrometer capable of making measurements of sufficient precision and accuracy. The basic instrument and the associated extraction

techniques, mostly from the 1950s, are still in use in many labs today. These techniques have become “conventional” in the sense of traditional, and they provide the benchmark against which the accuracy of techniques is compared. The 1986 volume was based almost exclusively on natural data obtained solely from conventional techniques. Since then, revolutionary changes in sample size, accuracy, and cost have resulted from advances in continues flow massspectrometry, laser heating, ion microprobes, and computer automation. The impact of new technology has differed by discipline. Some areas have benefited from vastly enlarged data sets, while others have capitalized on in situ analysis and/or micro- to nanogram size samples, and others have developed because formerly intractable samples can now be analyzed. Just as Stable Isotope Geochemistry is being reborn by new good ides, it is still being nurtured by new technology. The organization of the chapters in this book follows the didactic approach of the 2001 short course in Boston. The first three chapters present the principles and data base equilibrium isotope fractionation and for kinetic processes of exchange. Both inorganic and biological aspects are considered. The next chapter reviews isotope compositions through out the solar system including mass- independent fractionations that are increasingly being recognized on Earth. The fifth chapter covers the primitive composition of the mantle and subtle variations found in basalts. This is followed by three chapters on metamorphism, isotope thermometry, fluid flow, and hydrothermal alternation. The next chapter considers water cycling in the atmosphere and the ice record. And finally, there are four chapters on the carbon cycle, the sulfur cycle, organic isotope geochemistry and extinctions in geochemical record. The editors thank the authors of individual chapters and also all those who generously assisted in the scientific review of these chapters: J.Alt, R.Bidigare, I.Bindeman, J.Bischoff, J.Bowman, P.Brown, D.Canfield, W.Casey, F.Corsetti, J.Eiler, J.Farquhar, J.Farver, V.Fereirra, H.Fricke, M.Gibbs, B.Hanson, R.Harmon, J.Hayes, M.Hendericks, J.Horita, C.Johnson , Y.Katzir, E.King A.Knoll, J.S.Lackey, C.Lesher, K.Macleod, P.Meyers, W.Peck, B.Putlitz, F.Robert , G.Roselle, J.Severinghaus, Z.Sharp, M.Spicuzza, M.Thiemens, D.Tinker, and K.von Damn. Alex Speer and the MSA Business office Assisted with many aspects of the Short Course in Boston, before the November 2001 Meeting of the Geological Society of America. We thank the U.S. Department of Energy for providing publication support for this book and scholarships for students who attended the Short Course. We are especially indebted to Paul H Ribbe, series editor of the Reviews in Mineralogy and Geochemistry, formerly, Reviews in Mineralogy. Paul was responsible for general editing, final composition of the text, and motivating us to meet the production deadlines. Paul was also the series editor for the 1986 Stable Isotopes volume and has produced a remarkable total of more than forty RiM or RiMG volumes. This is a major achievement in affordable publishing of forefront science. John W. Valley Madison, Wisconsin

David R. Cole Oak Ridge, Tennessee August 15, 2001

This page left blank.

1

Equilibrium Oxygen, Hydrogen and Carbon Isotope Fractionation Factors Applicable to Geologic Systems Thomas Chacko Department of Earth and Atmospheric Sciences University of Alberta Edmonton, Alberta T6G 2E3, Canada

David R. Cole and Juske Horita Chemical and Analytical Sciences Division Oak Ridge National Laboratory Oak Ridge, Tennessee 37831 INTRODUCTION As demonstrated by the chapters in this short course, stable isotope techniques are an important tool in almost every branch of the earth sciences. Central to many of these applications is a quantitative understanding of equilibrium isotope partitioning between substances. Indeed, it was Harold Urey’s (1947) thermodynamically based estimate of the 18 16 temperature-dependence of O/ O fractionation between calcium carbonate and water, and a recognition of how this information might be used to determine the temperatures of ancient oceans, that launched the science of stable isotope geochemistry. The approach pioneered by Urey has since been used to estimate temperatures for a wide range of geological processes (e.g. Emiliani 1955; Anderson et al. 1971; Clayton 1986; Valley, this volume). In addition to their geothermometric applications, equilibrium fractionation data are also important in the study of fluid-rock interactions, including those associated with diagenetic, hydrothermal, and metamorphic processes (Baumgartner and Valley, this volume; Shanks, this volume). Finally, a knowledge of equilibrium fractionation is a necessary first step in evaluating isotopic disequilibrium, a widespread phenomenon that is increasingly being used to study temporal relationships in geological systems (Cole and Chakraborty, this volume). In the fifty-four years since the publication of Urey’s paper, equilibrium fractionation data have been reported for many minerals and fluids of geological interest. These data were derived from: (1) theoretical calculations following the methods developed by Urey (1947) and Bigeleisen and Mayer (1947); (2) direct laboratory experiments; (3) semi-empirical bond-strength models; and (4) measurement of fractionations in natural samples. Each of these methods has its advantages and disadvantages. However, the availability of a variety of methods for calibrating fractionation factors has led to a plethora of calibrations, not all of which are in agreement. In this chapter, we evaluate the major methods for determining fractionation factors. We also compile data on oxygen, hydrogen, and carbon isotope fractionation factors for geologically relevant mineral and fluid systems. Our compilation focuses primarily on experimental and natural sample calibrations of fractionations factors as large compilations of theoretical (Richet et al. 1977; Kieffer 1982) and bond-strength (e.g. Hoffbauer et al. 1994; Zheng 1999a) calibrations already exist in the literature. The chapter begins with a general overview of the theoretical basis of stable isotope fractionation, and theoretical methods for calculating fractionation factors. The reader is referred to the earlier review papers of Richet et al. (1977), Clayton (1981), O’Neil (1986) and Kyser (1987), and the recent textbook by Criss (1999) for more detailed

1529-6466/00/0043-0001$10.00

2

Chacko, Cole & Horita

discussion of theoretical topics. Our emphasis will be on advances in the determination of fractionation factors and on our understanding of the variables that control isotopic fractionation behavior made since the publication of Reviews in Mineralogy, Volume 16 (Valley et al. 1986). THEORETICAL BACKGROUND Comparison of cation and isotope exchange reactions The equilibrium fractionation of isotopes between substances is analogous in many ways to the partitioning of cations (such as Fe and Mg) between minerals. Both processes can be described in terms of chemical reactions in which isotopes or cations are exchanged between two coexisting phases. For example, the partitioning of Fe and Mg between orthopyroxene and olivine can be described by the reaction: FeSiO3 + 1/2 Mg2SiO4 = MgSiO3 + 1/2 Fe2SiO4 where the equilibrium constant, K1, for this reaction is: ol (a opx MgSiO3 )(a Fe2SiO4 ) K1 = opx 1/2 (a FeSiO3 ) (a ol Mg 2SiO4 ) 1/2

where aip is the thermodynamic activity of component i in phase p. Assuming ideal mixing of Fe and Mg on the octahedral sites of orthopyroxene and olivine, K1 can be recast as:

K1 =

(Mg/Fe)opx (Mg/Fe)opx 1/2 = [(Mg/Fe)2ol ] (Mg/Fe)ol

Similarly, the partitioning of 18O and 16O between olivine and orthopyroxene is described by the reaction: 1/3 MgSi16O3 + 1/4 Mg2Si18O4 = 1/3 MgSi18O3 + 1/4 Mg2Si16O4 Assuming ideal mixing of oxygen isotopes among the different oxygen sites in olivine and orthopyoxene, this gives an equilibrium constant, K2, of: [(18 O/16 O)3opx ]1/3 (18 O/ 16O)opx K 2 = 18 16 4 1/4 = 18 16 = α opx −ol [( O/ O)ol ] ( O/ O)ol If the reaction is written such that one mole of 18O and 16O atoms are exchanged between the two minerals, K2 is equal to αopx-ol, the oxygen isotope fractionation factor between orthopyroxene and olivine. As with all chemical reactions, the standard state Gibbs free energy change for an isotope exchange reaction at a given pressure and temperature is related to the equilibrium constant by: ∆GoR(T, P) = ∆HoR - T∆SoR + P∆VoR = -RT ln K

(1)

In principle, the free energy change and in turn the equilibrium constant for such reactions can be calculated from conventional thermodynamic data (molar enthalpy, entropy, volume data) on the end-member isotopic species denoted in the reaction. This approach, however, is generally not practicable because of the paucity of thermodynamic data on isotopically ‘pure’ end-members. Moreover, even if such data were widely available, the Gibbs free energy changes associated with most isotope exchange reactions

Equilibrium Isotopic Fractionation Factors

3

are too small (typically a few tens of joules or less, compared to thousands of joules for cation exchange reactions) to permit precise calculations using classical thermodynamic methods. Despite its limitations, the discussion above is useful for illustrating the formal similarities between cation and isotope exchange reactions. Equation (1) also shows that equilibrium constants for all exchange reactions are dependent on temperature. More specifically, ln K varies linearly with T-1 if ∆GoR is independent of temperature. This T-1 temperature-dependency generally applies to cation exchange reactions because values of ∆GoR for such reactions are approximately constant over a wide range of temperatures. In the case of isotope exchange reactions, however, ∆GoR varies significantly with temperature, which results in higher order temperature-dependencies (T-2). The effect of pressure on ln K is determined by the volume change for the reaction. For most isotope exchange reactions, ∆VoR is small, resulting in correspondingly small pressure effects on ln K. However, as discussed further below, pressure effects can be significant for hydrogen isotope fractionations, particular those involving water. QUANTUM MECHANICAL REASONS FOR ISOTOPIC FRACTIONATION

The existence of small but significant free energy changes in isotope exchange reactions implies energetic differences between chemical species differing only in their isotopic composition. These energy differences are entirely a quantum mechanical phenomenon arising from the effect of atomic mass on the vibrational energy of molecules. Consider a diatomic molecule, which can be represented by two masses, m1 and m2, attached by a spring (Fig. 1a). The force (F) exerted on the masses is equal to the displacement (x) of the spring from the rest position times the force constant (k) of the spring (i.e. the spring’s stiffness): F = -kx, The potential energy, PE, of the spring is given by the equation: PE = kx2/2 which defines a parabola with a minimum potential energy when the spring is at the rest position (x = 0), and increasing potential energy when the spring is compressed (-x) or stretched (+x) from that position (Fig. 1b). The vibrational frequency, ν, of the spring is given by:

ν=

1 2π

k

µ

(2)

where µ is the reduced mass and given by: m1 m2 µ= m1 + m2 Derivation of the equations given above can be found in most introductory physics or physical chemistry textbooks. McMillan (1985) also provides a helpful summary. In its simplest form, the spring-chemical bond analogy is referred to as the harmonic oscillator approximation. Several facets of this analogy are useful to keep in mind. Firstly, the rest position of the spring corresponds to the optimal distance between the nuclei of the two atoms, and the minimum in the potential energy curve. If the atoms are pushed closer or pulled further away than this optimal distance, electrical forces act to

4

Chacko, Cole & Horita

Figure 1. (A) Drawing of a spring with two masses attached (simple harmonic oscillator), which is an analogue for a diatomic molecule. (B) Schematic plot showing variation in the potential energy of harmonic (and anharmonic) oscillators as they are displaced from the rest position (x = 0). Energy levels are given by values of n. The zero point energy (ZPE) is the difference in energy between the bottom of the potential energy well and the energy of the ground vibrational state (n = 0). Note that the zero point energy of a molecule made with the heavy isotope (ZPE*) is lower than that of the molecule made with the light isotope. The magnitude of ∆ZPE (ZPE-ZPE*) for a substance exerts a major control on its isotopic fractionation behavior.

restore the atoms towards the equilibrium position. Secondly, a strong chemical bond can be thought of as a stiff spring (i.e. a spring with a large force constant). It follows from Equation (2) that, all else being equal, strong bonds generally have higher vibrational frequencies than weak bonds. Thirdly, according to the Born-Oppenheimer approximation, isotopic substitution has no effect on the force constant of a bond. The magnitude of the force constant is determined by the electronic interaction between atoms, and is independent of the masses of the two nuclei. Thus, the potential energy curves of molecules comprising heavy and light isotopes of an element are identical. Given the last statement, classical mechanics predicts no energy differences between two molecules that differ only in their isotopic composition. At a temperature of absolute zero, both molecules should have energies corresponding to the bottom of their identical potential energy wells. Quantum theory, however, indicates that the vibrational energy, E, is quantized and given by: E = (n + 1/2)hν

(3)

where n corresponds to the energy levels 0, 1, 2, 3, etc., and h is Planck’s constant. Thus, even at absolute zero, where all molecules are in the ground state (n = 0), the vibrational energy of these molecules lies some distance above the bottom of the potential energy well. The energy difference between the bottom of the potential energy well and the

Equilibrium Isotopic Fractionation Factors

5

energy of ground vibrational state is referred to as the zero point energy or ZPE (Fig. 1b). Importantly, although the potential energy curves of molecules made up of light and heavy isotopes of an element are identical, their ZPE’s are different because of the effect of mass on vibrational frequency. More specifically, it can readily be shown from Equation (2) that the ratio of the vibrational frequencies of isotopically heavy and light molecules of a particular compound is given by:

ν* µ = ν µ*

(4)

where the asterisks denote the molecule containing the heavy isotope. Because of the inverse relationship between frequency and mass, the heavy molecule has a lower vibrational frequency, and hence a lower ZPE than the light molecule (Fig. 1b). This implies that a isotopically heavy molecule is always energetically more stable than its isotopically light counterpart. It should be clear from the discussion above that all substances will be stabilized by heavy isotope substitution, and thus prefer to form bonds with the heavy isotope. The key issue for partitioning of isotopes between substances is the preference of one substance over another for the heavy isotope. This is determined by the degree to which a molecule’s vibrational energy is lowered by heavy isotope substitution. At low temperatures, where all molecules are in their ground state, the magnitude of energy lowering is, to a good approximation, given by:

∆ZPE = ZPE – ZPE* = 1/2 h(ν – ν*) = 1/2 h∆ν

(5)

In a competition for the heavy isotope, the substance with the larger ∆ZPE (Fig. 1b) is more stabilized by the isotopic substitution, and therefore takes the lion’s share of the heavy isotope. It should be noted that with increasing temperature, a progressively larger fraction of molecules are excited to higher energy levels (n > 0). In those cases, ∆ZPE remains an important factor, but not the only factor in determining isotope fractionation behavior. Earlier in this section, we stated that strong bonds tend to have higher vibrational frequencies than weak bonds. By rearranging terms in equation (4), it can be shown that, other things being equal, bonds with high vibrational frequency undergo larger frequency shifts (∆ν) on isotope substitution than bonds with low vibrational frequency.

 µ  ∆ ν = ν – ν * = ν 1 − µ* 

(6)

From Equation (5), it follows that large frequency shifts lead to large ∆ZPE, and consequently an affinity for the heavy isotope. The important generalization that stems from Equations (5) and (6) is that the heavy isotope favors substances with strong bonds. An example of this correlation between bond strength and heavy isotope partitioning is the sequence of 18O enrichment observed in coexisting silicate minerals. Taylor and Epstein (1962) noted that minerals with abundant Si-O bonds are enriched in the heavy isotope of oxygen (18O) relative to minerals with fewer Si-O bonds. This correlation reflects the high strength and therefore high vibrational frequency of Si-O bonds relative to other cation-oxygen bonds in silicates, and the effect of these parameters on ∆ZPE. The discussion above is based on the harmonic oscillator approximation, which is the simplest model for representing the energetics of a molecule. In this model, the potential energy curve is symmetrical and the energy levels are equally spaced. More

6

Chacko, Cole & Horita

realistic models, such as that of Morse (1929), have asymmetric potential energy curves, non-uniformly spaced energy levels, and numerically different values of ZPE than given in the harmonic model. Despite these differences, the general principles outlined in this section also apply to the more complex models. CALCULATING FRACTIONATION FACTORS Theory The detailed calculation of isotopic fractionation factors follows the approach of Urey (1947) and Bigeleisen and Mayer (1947), and the reader is referred to those papers for further explanation of the equations given below. A summary of the nomenclature used in these equations is given in Appendix 1. The equations were originally derived for ideal gases, and require additional approximations if applied to liquids or solids. The calculation of fractionation factors involves the partition function, Q, a statistical mechanical parameter that describes all possible energy states of a substance. The equilibrium constant for an isotope exchange reaction can be expressed as a ratio of the partition functions of the two sides of the reaction. For example, consider a generalized isotope exchange reaction between substances A and B: aA + bB* = aA* + bB where a and b represent stoichiometric coefficients, and the asterisk, here and in all subsequent references, denotes the substance made with the heavy isotope. The equilibrium constant, KA-B, for this reaction can be expressed as: KA-B =

(Q*)aA (Q) bB (Q * /Q)aA a b = b (Q)A (Q*) B (Q * /Q)B

(7)

To a good approximation, the total partition function (Q) for each species in the reaction is the product of the translational (tr), rotational (rot), and vibrational (vib) partition functions: Q = Qtr x Qrot x Qvib

(8)

Taking these partition functions individually, the translational partition function is given by:

Qtr =

(2π Mkb T) 3/2 h3

V

where M is the molecular weight, kb is Boltzmann’s constant and V is the volume of the system. Fortunately, the partition function of each species in the reaction need not be evaluated in the calculation of the equilibrium constant, only the ratio of partition functions of a species and its isotopically substituted derivative (e.g. [Q*]A/[Q]A). Because all ideal gases occupy the same volume at a given pressure and temperature, all terms except molecular weight cancel in the ratio of translational partition functions: ⎛ M *⎞ (Q * /Q)tr = ⎝ M ⎠

3/ 2

Similarly, most terms cancel in the calculation of the ratio of rotational partition functions. For diatomic molecules and linear polyatomic molecules, this ratio is given by: σI * (Q * /Q)rot = σ *I

Equilibrium Isotopic Fractionation Factors

7

where I is the moment of inertia and σ is the symmetry number, which is the number of equivalent ways of orienting a molecule in space. For example, σ = 1 for heteronuclear diatomic molecules (e.g. NO or HD), and σ = 2 for homonuclear diatomic molecules (e.g. O2). The rotational partition function ratio for non-linear polyatomic molecules is:

σ  I A * I B * I C * (Q * /Q) = σ *  IA I BIC 

1/2

rot

where IA, IB, and IC are the three principal moments of inertia. In the harmonic oscillator approximation, the vibrational partition function ratio is given by: -Ui */2 −Ui (Q * /Q)vib = ∏ e -Ui /2 1− e−Ui * e 1−e i

where Ui = hνi/kbT and i is a running index of vibrational modes. There is only one vibrational mode for diatomic molecules (i = 1). For linear and non-linear polyatomic molecules consisting of s atoms, there are 3s-5 and 3s-6 vibrational modes, respectively, all of which must be considered in the calculation of Q*/Q. Combining the contributions of translational, rotational and vibrational partition functions yields: (Q * /Q) =

( ) M* M

3/2

σ I * e -U */2 1 − e −U σ * I e-U /2 1 − e − U * i

i

i

i

for diatomic molecules and:

( )

M* (Q * /Q) = M

3/2

σ  I A * IB * IC *  σ *  I AI BI C 

1/2

-U i */2

∏ ee i

-Ui/2

1 − e − Ui − Ui * 1− e

for polyatomic molecules. The moments of inertia can be removed from the expressions through use of the Teller-Redlich spectroscopic theorem (Urey 1947). This yields:

( )

m* (Q * /Q) = m

3r/2

σ ν i* e-U */2 1 − e − U ∏ σ * i ν i e -U /2 1− e − U * i

i

i

i

(9)

where m* and m are the atomic weights of the isotopes being exchanged, and r is the number of atoms of the element being exchanged present in the molecule (e.g. r = 1 for oxygen exchange in CO; r = 2 for oxygen exchange in CO2). Equation (9) forms the basis for the calculation of fractionation factors for gaseous substances. Several features of Equation (9) are noteworthy (cf. Richet et al. 1977; Criss 1999). The first three terms on the right hand side of the equation ([m*/m]3r/2, [σ/σ*], [ν∗/ν]) which take into account the effect of translation and rotation on (Q*/Q), are independent * of temperature. The fourth term (e-Ui /2/e-Ui/2) varies with temperature, but is mainly controlled by the ZPE difference of the isotopically heavy and light molecules. The last * term ([1-e-Ui]/[1-e-Ui ]) relates to the spacing of energy levels. At low temperatures, where nearly all molecules are in the ground vibrational state, this term is close to unity, and therefore does not contribute appreciably to Q*/Q. The term has a progressively larger effect on Q*/Q as temperature increases. Finally, the mass term ([m*/m]3r/2) in Equation (9) cancels in the calculation of an equilibrium constant ([Q*/Q]A/[Q*/Q]B). That is, the mass term for one molecule (A) taken to the stoichiometrically appropriate

8

Chacko, Cole & Horita

exponent equals that for the other molecule (B) involved in the exchange reaction. Thus, the mass term need not be considered for our purposes. By convention, partition function ratios with the mass term omitted are called reduced partition function ratios, and sometimes referred to by the symbol f. f is formally defined as: Q *  m  3r/2 (10) Q  m * In tabulations of partition function calculations (e.g. Richet et al. 1977), reduced partition ratios are commonly reported as f 1/r = β values, or 1000 ln β values. In such cases, the fractionation factor between two substances is simply: f =

α(A-B) = βA/βB and 1000 ln α(A-B)= 1000 ln βA – 1000 ln βB The input data required in calculating fractionation factors are the vibrational frequencies of all chemical species participating in an isotope exchange reaction. In many cases, however, frequencies have only been measured for molecules made with the abundant isotope (e.g. 16O); the frequencies of molecules containing the rare isotope (e.g. 18 O) must be calculated. The simplest way to calculate the unknown frequencies is through the harmonic oscillator approximation (Eqn. 4). More rigorous and accurate calculations of frequencies require force-field models, which are available for many common gaseous molecules (e.g. Richet et al. 1977). An example calculation

As an example, we show the calculation of the 18O/16O fractionation factor between CO2 and CO. Such calculations were computationally laborious in Urey’s time but today can readily be done on a spreadsheet. The exchange reaction controlling oxygen isotope fractionation in the CO2-CO system is: C18O + 1/2 C16O2 = C16O + 1/2 C18O2 with an equilibrium constant given by: 1/2

K CO

2

− CO

=

QC18 O 2 QC 16 O 1/2

QC16 O 2 QC 18 O

=

(Q * /Q)1/2 CO 2 = α (CO2 (Q * /Q)CO

− CO)

For isotope exchange reactions written as above involving only isotopically pure molecules (e.g. pure C16O or C18O), the symmetry number of a molecule and its isotopic derivative are identical. Therefore, σ/σ∗ = 1, and the term need not be included in the calculations. The vibrational frequencies used in our calculations are the same as those on which Urey’s (1947) calculations are based. However, Urey corrected these frequencies for anharmonicity (zero-order frequencies), whereas we used observed (measured) fundamental frequencies with no anharmonicity correction (see discussions in Bottinga 1969a, p. 52; McMillan 1985, p. 15; and Polyakov and Kharlashina 1995, p. 2568). Vibrational frequencies are generally reported in wave numbers (ω), which have units of cm-1. For partition function calculations, wave numbers must be converted to units of sec-1 by multiplying by c, the velocity of light (ν = cω). There is one vibrational mode for diatomic molecules such as CO, and four (3s-5) vibrational modes for linear tri-atomic molecules such as CO2. The four modes of CO2 correspond to different vibrational motions of the CO2 molecule, the symmetric stretching vibration (ω1), the asymmetric stretching vibration (ω3), and two lowerfrequency bending vibrations (ω2) (Fig. 2). The two bending modes are referred to as degenerate because they have the same vibrational frequency. Therefore, although it is

Equilibrium Isotopic Fractionation Factors

9

Figure 2. Vibrational modes of the CO2 molecule, the symmetric stretching vibration (ω1), the asymmetric stretching vibration (ω3), and the two bending vibrations (ω2). Note that ω1 only involves movement of oxygen atoms, whereas ω2 and ω3 involve movement of both oxygen and carbon atoms. This results in a larger 18O frequency shift (∆ν) for the ω1 vibration.

listed only once in Table 1, ω2 must be counted twice in calculating the partition function ratio of CO2 using Equation (9). Note also that the magnitude of frequency shifts (∆ν) for the isotopically substituted CO2 molecule varies with vibrational mode (Table 1). The largest shift is associated with the ω1 vibration, which relates to the fact that this vibrational mode involves only movement of oxygen atoms, whereas the other three modes involve the movement of both oxygen and carbon atoms (Fig. 2). The reduced mass of the 12C16O2 molecule undergoing the ω1, ω2 and ω3 vibrations is then given by:

µω = 1

mo mo mo + mo

µω

2,ω 3

=

2mom c 2mo + m c

(see Polyakov and Kharlashina 1995) where mo and mc are the masses of the 16O and 12C atoms, respectively. As a result of these relationships, the change in reduced mass on 18O substitution ([µ/µ*]1/2), and therefore the frequency shift, is significantly greater for the ω1 vibration than for the other vibrational modes. Table 1 shows the contribution of individual terms to the total and reduced partition function ratios of CO and CO2. The dominant contributor to the reduced partition function ratios of both molecules is the ZPE term, particularly at low temperatures. Thus, as noted above, isotope partitioning between substances is strongly influenced by the magnitude of ∆ZPE (cf. Bigeleisen 1965). The large frequency shift associated with the ω1 vibration of the CO2 molecule results in a large ∆ZPE, and therefore a tendency for CO2 to concentrate the heavy isotope at low temperature. At higher temperature, the magnitudes of vibrational frequencies (high ν) play an increasingly important role in isotope partitioning. In the case of the CO2-CO system, the high vibrational frequency of the CO molecule results in a so-called crossover in fractionations between 0 and 500°C (crossover occurs at 289°C). That is, at T > 289°C, 18O partitions into CO rather than into CO2. Such crossovers are entirely consistent with equilibrium fractionation of isotopes, and, as noted by Urey (1947), Stern et al. (1968) and Spindel et al. (1970), are not uncommon in gaseous substances. The calculations above are made on the basis of the harmonic oscillator approximation, and include no explicit corrections for anharmonicity. The effects of anharmonicity can be incorporated by using calculated zero-order frequencies rather than observed fundamental frequencies (see above), and by adding anharmonic corrections to the ZPE and energy level spacing terms of Equation (9) (Urey 1947; Bottinga 1968; Richet et al. 1977). Urey (1947) included anharmonic effects in his calculations of partition function

ω1 ω2 (2) ω3 500 500 500 500

0 0 0 0

0 500

T(°C)

1345.4 661.7 2349.5

1268.3 651.6 2313.5

C18O2 1268.3 651.6 2313.5

12

12

C16O2 1345.4 661.7 2349.5

C18O 2089.1 2089.1

12

12

C16O 2140.8 2140.8

ω* (cm-1)

(1)

ω (cm-1)

(1)

(m*/m)3r/2

1.424987 1.424987 1.424987

1.424987 1.424987 1.424987

1.193728 1.193728

(2)

1.074391 1.009468 1.034025

1.225194 1.027031 1.099334

1.145914 1.049296

e-U*/2/e-U/2

K (0°C) = 1.013203 K (500°C) = 0.999033

0.942681 0.984695 0.984695

0.942681 0.984695 0.984695

0.975843 0.975843

ν*/ν

1.496899

1.829239

1.334870 1.224662

Q*/Q

1.024922

1.133000

1.118236 1.025914

f =β

(4) 1/r

1000lnα (0°C) = 13.12 ‰ 1000lnα (500°C) = -0.97 ‰

1.013936 1.007904 1.000886

1.000420 1.001735 1.000001

1.000004 1.001920

(1-e-U)/ (1-e-U*)

(4) f is the reduced partition function ratio where f = (Q*/Q) (m/m*)3r/2. The equilibrium constants listed above are for the isotope exchange reaction given in the text; K (CO2-CO) = α (CO2-CO) = {(Q*/Q)0.5CO2/(Q*/Q)CO} = βCO2/βCO.

(3) ω2 for the CO2 molecule is degenerate. That is, two vibrational modes of the molecule have this frequency.

(2) The term r in the exponent represents the number of atoms of the isotope being exchanged present in the molecule (e.g., r = 1 for CO and 2 for CO2).

(1) Observed fundamental vibrational frequencies for CO and CO2 corresponding to zero-order frequencies (corrected for anharmonicity) given in Urey (1947). The vibrational frequencies of the C18O and C18O2 molecules (ω*) were calculated using Equation (4) (see text).

(3)

CO2 ω1 (3) ω2 (2) ω3

CO ωθ ωθ

Vibrat. mode

Table 1. Calculation of 18O/16O fractionation factors between CO2 and CO.

10 Chacko, Cole & Horita

Equilibrium Isotopic Fractionation Factors

11

ratios for CO2 and CO. Because we used the same vibrational frequencies as Urey, his results in the CO2-CO system can be compared directly with those given in Table 1. For the five temperatures (273-600 K) for which Urey reported fractionation factors, our calculations neglecting anharmonicity are within 0.04 to 0.35‰ of his. Thus, the net effect of anharmonicity on fractionation factors in this system is relatively small, albeit in some cases outside measurement error. In their detailed partition function ratio calculations for gaseous molecules, Richet et al. (1977) showed that the largest anharmonic correction is to the ZPE term but that the magnitude of this correction decreases with increasing with temperature. The anharmonic correction to the energy level spacing term is much smaller but increases with increasing temperature. For most gaseous substances, the net effect of including the anharmonicity terms is to decrease the calculated β value (Urey 1947; Richet et al. 1977). Thus, in the calculation of fractionation factors from β values, the anharmonicity correction for one substance is often partly cancelled by the anharmonicity correction for the other substance in the exchange couple. More detailed discussions of anharmonicity are given in Bottinga (1968), Richet et al. (1977), Gillet et al. (1996) and Polyakov (1998). Calculation of fractionation factors for gases, liquids and fluids

Following the original compilation by Urey (1947), several studies have calculated fractionation factors involving geologically relevant gaseous molecules. The most sophisticated and widely cited of these studies is that of Richet et al. (1977), who reported oxygen, hydrogen, carbon, sulfur, nitrogen and chlorine isotope fractionation factors for a large number of gaseous species. Their calculations used the latest (at the time) spectroscopic data and theoretical models, and included anharmonicity terms for all the gaseous species considered. In general, their calculated fractionation factors are in good agreement with experimental data. Although the basic principles still hold, isotopic fractionation theory developed for ideal gases is not directly applicable to liquids. Indeed, it has long been known from experiments that gases fractionate isotopes relative to liquids of the same composition. For example, the equilibrium hydrogen and oxygen isotope fractionation between liquid and gaseous H2O is 73 and 9.2‰, respectively, at 25°C (Majoube 1971b; Horita and Wesolowski 1994). These large fractionations are the result of two effects in the condensed phase (Bigeleisen 1961; Van Hook 1975). First, translational and rotational energy levels, which for free moving gaseous molecules are well represented by an energy continuum, become quantized in liquids because of interactions among molecules. Thus, whereas there is no isotope fractionation associated solely with translation or rotation in gases, there is such a fractionation in the liquid phase. This effect favors partitioning of the heavy isotope into a liquid relative to a gas of the same composition. The second effect relates to the influence of intermolecular interactions on the vibrational characteristics of individual molecules in the liquid. Thus, vibrational frequencies of substance measured in the gas phase are not identical to those of the same substance in the liquid phase. This second effect may cause isotopic fractionation in the opposite direction to the first (Bigeleisen 1961), and lead to a crossover in liquid-vapor fractionations, as has been documented for D/H fractionations in the H2O(l)-H2O(v) system. We are not aware of any direct calculations of partition function ratios for supercritical fluids. However, experimentally determined oxygen and carbon isotope fractionations in the CO2-calcite system at supercritical conditions are in good agreement with fractionations derived from theoretical calculations (Chacko et al. 1991; Scheele and Hoefs 1992; Rosenbaum 1994). This suggests that supercritical CO2 is well represented by the ideal gas approximation as the calculations were made on this basis (Rosenbaum

12

Chacko, Cole & Horita

1997). The same conclusion does not appear to hold for H2O (Bottinga 1968; Clayton et al. 1989; Rosenbaum 1997; Driesner 1997). For example, oxygen partition function ratios for supercritical H2O derived empirically from mineral-H2O exchange experiments are distinctly lower than those calculated using the ideal gas approximation (Rosenbaum 1997). As first suggested by Bottinga (1968), this discrepancy may be due to the nonideality of H2O under supercritical conditions, and the effect of this non-ideality on the vibrational frequencies of the H2O molecule. Another possible reason for the discrepancy is the solubility of minerals in water at elevated pressures and temperatures. Thus, partition function ratios of H2O derived empirically from mineral-H2O experiments may be different from those of a pure H2O fluid at comparable P-T conditions (Hu and Clayton, in press). Calculation of fractionation factors for minerals Principles. Partition function ratios can also be calculated for minerals, but these calculations are complex and require a number of approximations. The general approach is to treat a mineral as a large molecule consisting of 3s independent oscillators, where s is the number of atoms in the unit cell. For example, quartz, which contains 9 atoms in its unit cell (Si3O6), has a total of 27 vibrational modes. Of these, 24 (3s-3) are so-called optical modes because their vibrational frequencies are derived from optical spectroscopic techniques (e.g. IR or Raman spectroscopy). The optical modes are subdivided into internal modes, which concern the vibrational motions of individual functional groups within the mineral (e.g. Si-O in silicates and CO3 in carbonates), and external modes, which correspond to vibrations of the mineral lattice. The remaining 3 modes are referred to as acoustic modes because they relate to sound velocity. The frequencies of the acoustic modes are typically derived from heat capacity and spectroscopic data using Debye-Einstein models (e.g. Bottinga 1968; Kawabe 1978). Using these data, the partition function ratio for the unit cell in the harmonic approximation is given by (Kieffer 1982): 3s

(Q * /Q) =

e -Ui */2 1 − e −Ui ∏ e-U i/2 1 − e − Ui* i= 1

(11)

Equation (11) for minerals is similar to Equation (9) for gases but ignores translational and rotational contributions to the partition function, as such motions are restricted or absent in solids (Bottinga 1968). In detailed calculations, the three acoustic modes are treated separately from the optical modes because each acoustic mode represents a continuous spectrum of vibrations rather than a single vibrational frequency. As such, the acoustic modes are best evaluated by means of Debye functions rather than the Einstein functions typically used to treat the optical modes (see Eqn. 6 in Bottinga (1968) for the mathematical details of dealing with the acoustic modes). The total partition function ratio given by Equation (11) can be converted to a reduced partition function ratio using Equation (10) (Kieffer 1982), but r in the case of minerals is the number of atoms being exchanged in the unit cell (e.g. r = 6 for oxygen exchange in quartz). The largest uncertainty in the calculation of partition function ratios for minerals is the magnitude of frequency shifts on isotope substitution. Because direct spectroscopic measurements of minerals made with the less abundant isotope are not widely available, these shifts must usually be calculated or estimated in some other way. The detailed approach to calculating frequency shifts employs the methods of lattice dynamics to determine force constants for each vibrational mode, from which the vibrational frequencies for a mineral and its isotopic derivative can be predicted (e.g. Bottinga 1968;

Equilibrium Isotopic Fractionation Factors

13

Elcombe and Hulston 1975; Kawabe 1978). Kieffer (1982) took a less detailed approach in calculating oxygen isotope partition function ratios for 11 silicate minerals, calcite and rutile. As input for her calculations, Kieffer used the measured spectra for the 16O forms of minerals, divided the vibrational modes for these minerals into four groups, and then developed a set of rules for estimating the frequency shift associated with each group on 18 O substitution. Importantly, she applied the same rules to each mineral considered in her study, which resulted in an internally consistent set of partition function ratios. Most of Kieffer’s calculated fractionation factors are in excellent agreement with experimental data (Clayton and Kieffer 1991). Table 2. Calculation of oxygen isotope partition function ratio for quartz at 25°C. ω (cm-1) a

102 122 a 164 128 205.6 263.1 354.3 363.5 393.8 401.8 450 463.6 509 697.4 796.7 808.6 1066.1 1083 1160.6 1231.9 a

(1)

shift factor 0.96522 0.96522 0.96522 0.93828 0.93677 0.95173 0.97883 0.97387 0.9647 0.95744 0.9426 0.95513 0.9426 0.98179 0.98795 0.98417 0.96398 0.96602 0.95614 0.965

(2)

gi

1 1 1 2 1 2 1 1 1 1 2 1 1 2 1 2 2 1 2 1

(3)

f(x)gi

Q*/Q

f

1000 lnβ

1.03675 1.03705 1.03786 1.14019 1.07279 1.11772 1.02662 1.03332 1.04688 1.05741 1.17246 1.06491 1.08850 1.06803 1.02450 1.06652 1.20647 1.09400 1.28109 1.11031

5.3344

1.8435

101.95

ln (Q*/Q)optical = 1.56455 ln (Q*/Q)acoustic = 0.10963 ln (Q*/Q)total = 1.67418

(1) Frequency shift factor (ω*/ω) (2) gi is the degeneracy of the ith vibrational mode (3) f(x)gi = [(e-Ui*/2)/(e-Ui/2)][(1-e-Ui)/(1-e-Ui*)]gi. Frequencies and frequency shift factors taken from compilation of Polyakov and Kharlashina (1994) except for the shift factors of the three acoustic modes (denoted by the (a) superscript), which were calculated using the high-temperature product rule (see text).

Example calculation for quartz. We show, as an example, the calculation of a partition function ratio for quartz at 25°C (Table 2). The input data are taken from the compilation of frequencies and frequency shifts factors (ω*/ω) given in Polyakov and Kharlashina (1994), with minor modifications (see below). The shift factors in this compilation are mostly from Sato and McMillan (1987), who directly measured the spectrum of 18O quartz. The degeneracy column in the table gives the number of vibrational modes with a particular frequency, and therefore the number of times that mode must be counted in the calculations. Including degeneracies, there are 27 vibrational modes.

14

Chacko, Cole & Horita

(Q*/Q) was calculated by substituting ω and ω* for each vibrational mode into Equation (11). For the sake of simplicity, all the vibrational modes, including the acoustic modes, were represented by Einstein functions. More rigorous calculations would treat the acoustic modes using Debye functions. Given a value for (Q*/Q), f and β values are given by: 9  m 16  1   1000 ln β = 1000 lnf f = (Q * /Q)  m 18  6 When divided into contributions of optical and acoustic modes, the three low frequency acoustic modes contribute only about 6% to the partition function ratio for quartz (Table 2). The percentage is similarly low or lower in most minerals (O’Neil 1986). Thus, imperfect information on the acoustic frequencies typically does not lead to large errors in calculated partition function ratios. The 1000 ln β value for quartz at 25°C given in Table 2 (101.95) can be compared to values of 102.04 and 104.54 calculated by Kieffer (1982; corrected for a rounding error by Clayton et al. 1989) and Clayton and Kieffer (1991), respectively. The 2.6‰ difference in the results of these calculations is primarily due to differences in the input vibrational frequency data, and illustrates the sensitivity of theoretical calculations to these input data. It should be noted, however, that same input data yield 1000 ln β values of 11.43 (this study), 11.55 (Kieffer 1982), and 11.71 (Clayton and Kieffer 1991) at 1000 K, a range of only 0.3‰. Thus, the absolute magnitude of the discrepancy decreases markedly with increasing temperature, a consequence of the way that uncertainties in the input data propagate with temperature in Equations (9) or (11) (Richet et al. 1977; Clayton and Kieffer 1991; Farquhar 1995). High-temperature product rule. An important consideration in partition function calculations for minerals is ensuring the proper high-temperature limiting behavior. That is, ln f should go to zero as temperature goes to infinity (see below). However, this requirement is not always met in the calculations because of rounding errors and uncertainties in frequency shift factors. The problem can be avoided through use of the high-temperature product rule (Becker 1971; Kieffer 1982; Chacko et al. 1991):

 ω  ∏   i= 1 ω * 3s

gi

 m * = m

3r/2

where gi is the degeneracy of a vibrational mode. If the product of the frequency shift factors for all vibrational modes does not equal the quantity on the right hand side of the equation, ln f will not go to zero at infinite temperature. To correct the problem, one or more of the frequency shift factors must be adjusted so as to satisfy the equation. Typically, it is the shift factors for the lower frequency modes that are modified (O’Neil et al. 1969; Kieffer 1982; Chacko et al. 1991). In Table 2, the shift factors for the acoustic modes of quartz were changed to fulfill the product rule. It should be emphasized that this procedure can be quantitatively important, and affects partition function ratios at both high and low temperature. In the case of the CO2-calcite system, calculations that utilized the product rule gave results in much better agreement with experimental data than those that did not (Chacko et al. 1991). Other theoretical methods. In addition to the procedure described above, three other theoretical methods have been developed for calculating fractionation factors involving minerals. The first is based on computer simulation of crystal structures and first principles prediction of their thermodynamic properties (Patel et al. 1991; Dove et al.

Equilibrium Isotopic Fractionation Factors

15

1994). Reduced partition function ratios for calcite and a number of silicate minerals calculated using this approach are within 13% of those calculated by traditional methods. Fractionation factors (at 1000 K) derived from these calculations are within 1‰ and, in many cases, within 0.5‰ of those given by experiments. These early results suggest that the ab initio approach to calculating fractionation factors is promising, and should be pursued. At present, however, the approach may not be sufficiently accurate to provide quantitatively reliable fractionation factors. The second of the alternative methods, which is based on thermodynamic perturbation theory, can be applied to single element substances such as graphite or diamond (Polyakov and Kharlashina 1995). The method uses only heat capacity data for the minerals of interest as input. Application of this method to the diamond-graphite and calcite-graphite systems yielded results in good agreement with more standard theoretical calculations (Bottinga 1969b), and natural sample data (Valley and O’Neil 1981; Kitchen and Valley 1995), respectively. The third method is also based on thermodynamic perturbation theory but uses Mössbauer data as input (Polyakov 1997; Polyakov and Mineev 2000). The method can be applied to twoelement compounds if one of the elements (e.g. Fe) has a Mössbauer-sensitive isotope. Polyakov and Mineev (2000) used this approach to calculate iron, sulfur, and oxygen isotope reduced partition function ratios for a number of minerals. Iron isotope fractionations derived with this approach are in agreement with fractionations calculated using more traditional theoretical methods and vibrational spectroscopic data (Schauble et al. 2001). VARIABLES INFLUENCING THE MAGNITUDE OF FRACTIONATION FACTORS Temperature

Temperature is in many cases the single most important variable in controlling isotope fractionation behavior. Bigeleisen and Mayer (1947), Urey (1947), Stern et al. (1968), Bottinga and Javoy (1973), and Criss (1991) evaluated the temperature-dependence of fractionation factors from a theoretical perspective. The following is based largely on the lucid explanation provided in Criss’ (1991) paper. Written in logarithmic form, the reduced partition function ratio of a diatomic gas (see Eqn. 9) is given by:

σ ν *  U − U *  1− e -U   ln f = ln + + ln  -U* σ * ν   2  1 - e 

(12) *

As noted previously, the quantity [(1–e-U)/(1–e-U )] is approximately unity at low temperature. Thus, at low temperature, Equation (12) reduces to:

σ ν *  U − U * ln f ≅ ln  + σ * ν   2 

(13)

Because U = hν/kbT, this equation is of the form y = constant + slope (T-1), with a slope given by ∆ZPE/kb. The equation indicates that reduced partition function ratios, and in turn fractionation factors (ln α), vary linearly with respect to T-1 at low temperature. The same relationship applies to polyatomic molecules but in that case both terms in Equation (13) are summations over all vibrational modes of the molecule. At high temperature, the last term on the right-hand side of Equation (12) is significantly greater than zero and must be considered in the calculation. Criss (1991) expanded this term in a Taylor series, which, after canceling like terms, gives the following equation for diatomic molecules:

16

Chacko, Cole & Horita

σ   U2 − U*2  U 4 − U*4   U6 − U*6   U8 − U*8   − + − + ... + ln f = ln  σ *   24   2880   181,440   9,676,800  The higher order terms (terms 3 and above on the right hand side) become vanishingly small at high temperature, which results in:

σ  U2 − U*2  (14) ln f ≅ ln   +   σ *   24  This equation is of the form y = constant + slope (T-2), and implies that fractionation factors vary linearly with respect to T-2 at high temperature. Note also that ln f → 0 as T → ∞ provided that the symmetry numbers of a molecule and its isotopic derivative (σ/σ∗) are the same. The theoretical considerations discussed above provide insight into the expected temperature-dependence of fractionation factors at low- and high-temperature limits. The exact temperatures at which these limits occur depend on the substance being considered. Bigeleisen and Mayer (1947) showed that the reduced partition function ratios approach the T-1 and T-2 temperature-dependencies when values of hcω/kbT are >20 and 20 kbar). In contrast, the pressure effect on carbon isotope fractionations involving graphite is significant even at high temperature.

Carbon isotope fractionations. The same conclusion does not hold for carbon isotope fractionation factors involving graphite. Polyakov and Kharlashina (1994) noted that the β values of graphite are much more strongly affected by pressure than those of calcite or diamond (Fig. 5). As a consequence, the pressure effect on carbon isotope fractionations in the diamond-graphite and calcite-graphite systems is significant, even at high temperature. The pressure effect on graphite is, in fact, large enough to induce a

20

Chacko, Cole & Horita

fractionation crossover in the diamond-graphite system, with graphite becoming the 13 C-enriched phase at high pressure. Hydrogen isotope fractionations. Several recent studies indicate that pressure is an important variable in the hydrogen isotope fractionation behavior of hydrous mineralH2O systems. Using spectroscopic data on high temperature and pressure H2O as input, Driesner (1997) calculated large pressure effects on the D/H reduced partition function ratio of water. The effects were largest at the critical temperature of water (374°C), where the calculations predict a 20‰ decrease in the reduced partition function ratio from 0.2 to 2 kbar. Driesner assumed that the effect of pressure on the reduced partition function ratios of hydrous minerals would be much smaller than on those of the water molecule. Therefore, most of the calculated shifts for water should translate to similar magnitude shifts in mineral-H2O fractionation factors. Horita et al. (1999) tested this hypothesis with experiments in the brucite-water system. At 380°C, they found a 12.4‰ increase in brucite-H2O fractionations associated with a pressure increase from 0.15 to 8 kbar (Fig. 6). This change in fractionation factor is well outside experimental error and represents the first unequivocal demonstration of a pressure effect on D/H fractionations. Significantly, more than one half of the total change in the fractionation factor occurs from 0.15 to 0.54 kbar, which is the region of P-T space characterized by the proportionately largest increase in the density of water. The direction of the pressure effect documented by Horita et al. (1999) is the same as that indicated by Driesner’s (1997) calculations, but its magnitude is markedly smaller. It must be emphasized, however, that pressure effects on brucite-water D/H fractionation are substantial, and in fact larger than temperature effects in this system over the temperature range from 200 to 600°C (Horita et al., in press) (Fig. 7).

Figure 6. Pressure effect on D/H fractionations in the brucite-H2O system at 380°C (after Horita et al. 1999). Note that increasing pressure decreases water’s affinity for deuterium. The largest change in the fractionation factors occurs below 0.5 kbar, which is the region of P-T space with the proportionately largest increase in the density of water.

Although experimental details are not given, Mineev and Grinenko (1996) also report significant pressure effects on D/H fractionations in the serpentine-H2O system at 100 and 200°C. They suggest that the large discrepancy between the experimental serpentine-water calibration of Sakai and Tatsumi (1978), and the natural sample calibration of Wenner and Taylor (1973) can be attributed to differences in pressure. Similarly, the differences in tourmaline-H2O D/H fractionations obtained by Blamart

Equilibrium Isotopic Fractionation Factors

21

Figure 7. Experimental results of brucite-water D/H fractionation as a function of temperature and pressure. The calculated curve (dashed) is based on theoretical calculations for brucite at 0 kbar (Horita et al., in press), and water (Richet et al. 1977). The calculated curve has large errors (±20‰ at 300°C to ±15‰ at 700°C). The experiments indicate that the effect of pressure on D/H fractionations in this system is larger than the effect of temperature from 200-500°C. Modified after Horita et al. (in press).

et al. (1989) and Jibao and Yaqian (1997) can in part be attributed to differences in experimental pressures. In all the cases investigated thus far, the effect of increasing pressure is to decrease the water molecule’s affinity of deuterium. Although the direction of the pressure effect seems well established from both theory and experiment, the magnitude of the effect and the relative contributions of mineral and water to that effect need to be evaluated with additional experiments and calculations. Mineral composition Even if isotope fractionation factors are well determined for the compositional endmember of a particular mineral, application of these data to the full range of geological samples commonly requires additional information on how compositional variation in that mineral or mineral group affects fractionation behavior. Compositional effects on fractionation factors have been investigated in a number of different ways including theoretical calculations, experiments, natural sample data, and bond-strength methods. We defer our discussion of bond-strength methods to a later section but see Zheng (1999b) for an application of this methodology to the assessment of fractionations in carbonate and sulfate minerals. Compositional effects in carbonates. The classic experimental study of O’Neil et al. (1969), which was later refined by Kim and O’Neil (1997), systematically investigated the effect of cation substitution on the oxygen isotope fractionation behavior of carbonate minerals. Figure 8 shows experimentally determined carbonate-H2O fractionation factors at 240°C plotted against the radius and mass of the divalent cation. Although there is an overall negative correlation between the fractionation factor and both of these variables, the correlation with mass is considerably better. This suggests that cation mass rather than radius is the major variable controlling fractionations between carbonates (O’Neil et

22

Chacko, Cole & Horita

Figure 8. The effect of cation substitution on carbonate-H2O fractionation factors at 240°C (O’Neil et al. 1969; Kim and O’Neil 1997). Note that the change in fractionation factor with cation substitution correlates with both cation radius and cation mass; however, the correlation with mass is considerably better suggesting that cation mass is dominant variable.

al. 1969; Kim and O’Neil 1997). Golyshev et al. (1981) reached the opposite conclusion on the basis of lattice dynamical calculations for a large number of carbonate minerals. It should be noted, however, that Golyshev et al.’s calculated ln β values are at least as well correlated with cation mass as with cation radius. Both O’Neil et al. and Golyshev et al. pointed out that cation radius mainly affects the internal frequencies of the CO3 ion, whereas cation mass affects the lattice vibrations (acoustic and external optical frequencies). Notably, the frequencies of the lattice vibrations are much more strongly modified by cation substitution than are the internal frequencies. Compositional effects in silicates. Compositional effects are also important in the oxygen isotope fractionation behavior of silicate minerals. The effects are complex, however, because of the large number of substitution mechanisms that operate in these minerals. Table 3 summarizes some of the major substitutions and their estimated effects on isotope fractionation (cf. Kohn and Valley 1998a). Of the common substitution schemes, the plagioclase substitution has the largest isotopic effect. Experimental data (O’Neil and Taylor 1967; Matsuhisa et al. 1979; Clayton et al. 1989) and theoretical calculations (Kieffer 1982) indicate a 1.05 to 1.2‰ fractionation between albite and anorthite at 1000 K. This fractionation reflects the higher Si to Al ratio of albite, and the 18 affinity of the Si-O bond relative to the Al-O bond for O (Taylor and Epstein 1962). In contrast to plagioclase, K ↔ Na substitution in the alkali feldspars, which does not affect the Si to Al ratio, has no measurable isotopic effect (Schwarcz 1966; O’Neil and Taylor 1967). Like the plagioclase substitution, the jadeite (NaAlSi2O6)-diopside (CaMgSi2O6) substitution also involves Al, but Al in this case substitutes into an octahedral rather than a tetrahedral site. As demonstrated by a comparison of experimental data in the diopside-H2O and jadeite-H2O systems (Matthews et al. 1983a), the isotopic effect of this substitution is an enrichment in 18O (∆(jd-di) = 0.99‰ at 1000 K). Another substitution involving Al, the Tschermak substitution ([AloctAltet] ↔ [M2+Si]), can be significant in pyroxenes, amphiboles and micas. Unfortunately, there are no data with which to evaluate its isotopic effect directly. A crude estimate can be obtained by combining the measured isotopic effects of the plagioclase and jadeite substitutions. The large 18O depletion associated with replacement of Si by Al in the tetrahedral site (plagioclase) 18 should be mostly cancelled by O enrichment associated with replacement of a divalent

Equilibrium Isotopic Fractionation Factors

23

Table 3. Effect of compositional substitutions on oxygen isotope fractionation in silicates. Substitution NaSi↔CaAl NaAl↔Ca(Mg,Fe) K↔Na Fe↔Mg Mn↔Ca (Fe,Mg)↔Ca (Fe,Mg)↔Ca Al3+↔Fe3+ F↔OH

Example plagioclase jadeite-diopside alkali feldspar pyroxene, garnet garnet pyroxene garnet garnet phlogopite

Experimental 1.05, 21.07, 31.09 5 0.99 2 0 5 0.08 8 0 9,10 0.49 ---12 1.3 13 0.52

1

Theoretical 4 1.2 ------7 0 ---4 0.4 4 0.18, 70.12 4 0.5, 70.45 ----

Natural ------6 0 ------11 0.55-0.75 11 0.5 11 0.7 ----

Notes: Substitution schemes are written such that the left-hand side has the greater affinity for 18O. Experimental, theoretical and natural (sample) refer to the methodology used to evaluate the magnitude of the fractionation factor. All fractionations are reported at 1000 K and refer to the isotopic fractionation between end-members. For example, the numbers listed for the plagioclase substitution represent the fractionation between end-member albite and anorthite at 1000 K. Sources of data: 1 = Clayton et al. (1989); 2 = O’Neil and Taylor, (1967); 3 = Matsuhisa et al. (1979); 4 = Kieffer (1982); 5 = Matthews et al. (1983a); 6 = Schwarcz (1966); 7 = Rosenbaum and Mattey (1995); 8 = Lichtenstein and Hoernes (1992); 9 = Chiba et al. (1989); 10 = Rosenbaum et al. (1994); 11 = Kohn and Valley (1998b); 12 = Taylor (1976); 13 = Chacko et al. (1996).

cation by Al in the octahedral site (jadeite). Therefore, the net effect of Tschermak substitution is a relatively small depletion in 18O. Following the approach of Kohn and Valley (1998a), we estimate the magnitude of that depletion at 1000 K to be ~0.1‰ for amphiboles, ~0.2‰ for micas, and ~0.4‰ for pyroxenes per mole of Al substitution in the tetrahedral site. 2+

2+

Fe -Mg substitutions are common in many silicate minerals. Experimental data on calcic clinopyroxenes (Ca[Fe,Mg]Si2O6) at 700°C indicate no significant difference in mineral-H2O fractionation factors between Fe and Mg end-members (Matthews et al. 1983a). Similarly, calculated reduced partition function ratios of pyrope (Mg3Al2Si3O12) and almandine (Fe3Al2Si3O12) garnet are identical (Rosenbaum and Mattey 1995). Collectively, these observations suggest that the isotopic effect of Fe-Mg substitution in silicates is negligible, at least at high temperature. The same may be true for Ca-Mn substitutions as mineral-H2O experiments with grossular (Ca3Al2Si3O12) and spessartine (Mn3Al2Si3O12) garnets gave identical fractionations at 750°C (Lichtenstein and Hoernes 1992). In contrast to Ca-Mn, there does appear to be a small but significant isotopic effect associated with Ca-Mg and Ca-Fe2+ substitutions. Theoretical calculations (Kieffer 1982), and data from experiments (Chiba et al. 1989; Rosenbaum et al. 1994) and natural samples (Kohn and Valley 1998b) indicate fractionations of 0.4, 0.5 and 0.55-0.75‰, respectively, between orthopyroxene ([Fe,Mg]2Si2O6) and calcic clinopyroxene at 1000 K. Similarly, calculations and natural sample data suggest a 0.1 to 0.5‰ fractionation between pyrope-almandine and grossular garnet at 1000 K (Kieffer 1982; Rosenbaum and Mattey 1995; Kohn and Valley 1998b). There is also a significant isotopic effect associated with the Fe3+-Al3+ substitution mechanism. Taylor (1976) reported a 1.7‰ fractionation between grossular and andradite 3+ (Ca3Fe 2Si3O12) garnet in hydrothermal experiments at 600°C, which translates to a 1.3‰ fractionation at 1000 K. Theoretical calculations (Kieffer 1982; Rosenbaum and Mattey 1995), and natural sample data (Kohn and Valley 1998b) indicate smaller

24

Chacko, Cole & Horita

fractionation of 0.5 to 0.8‰ at that temperature. The generalizations that stem from the above observations are similar to the ones made long ago by Taylor and Epstein (1962). Namely, the dominant compositional variable affecting oxygen isotope fractionations between silicates is the identity of the tetrahedral cation. With the exception of Al, the octahedral and cubic (8-fold) cations are of secondary importance, although, as shown in Table 3, not insignificant in all cases. Monovalent cations have little effect on oxygen isotope fractionations. There is good overall agreement between theory, experiment and natural sample data as regards the direction, and, in some cases, the magnitude of isotopic effects associated with compositional substitutions in silicates. This agreement is encouraging because it suggests that fractionation factors for silicate solid solutions can be predicted by combining fractionation data for end-members with the estimated isotopic effects of the various substitution mechanisms. The reader is referred to the paper of Kohn and Valley (1998a) for an elegant methodology for making such calculations. Compositional effects on hydrogen isotope fractionation. As was the case with oxygen, hydrogen isotope fractionation factors are also influenced by mineral composition. The exact nature of the compositional dependence, however, remains unclear. From the results of their pioneering exchange experiments between micas, amphiboles and water, Suzuoki and Epstein (1976) concluded that the identity of the octahedral cation is the key compositional variable in the hydrogen isotope fractionation behavior of hydrous minerals. This observation can be rationalized by noting that the hydroxyl unit is more closely associated with the octahedral cations than with other cations in these mineral structures. Of the common octahedral cations, Al has the greatest affinity for deuterium, followed closely by Mg. Fe, on the other hand, has a strong affinity for hydrogen over deuterium. Suzuoki and Epstein (1976) suggested that these compositional effects are systematic and might be used to predict fractionation factors regardless of mineral species. Their proposed equation, which is applicable from 400 to 850°C, is: 1000 ln α (mineral-H2O) = -22.4 (106T-2) + 28.2 +(2 XAl – 4 XMg – 68 XFe) where XAl, XMg and XFe are the mole fractions of each cation in the octahedral site. This equation correctly predicts the magnitude (but not the temperature-dependence) of brucite-H2O fractionation factor (Satake and Matsuo 1984; Horita et al. 1999) to within 10‰ at 400 and 500°C. It is also supported in a general way by natural sample data, which suggest a negative correlation between the δD values of minerals and their Fe/Mg ratio (e.g. Marumo et al. 1980). There are, however, a number of complicating factors. Minerals such as boehmite, epidote and chlorite, which show a significant degree of hydrogen bonding (i.e. hydrogen exists as O-H--O units rather than O-H units), partition deuterium less strongly than predicted by the equation (Suzuoki and Epstein 1976; Graham et al. 1980, 1987). Additionally, a detailed study of hydrogen isotope fractionations between amphiboles and water does not show the compositional dependence indicated by Suzuoki and Epstein’s equation, and suggests that the A-site cation in amphiboles may also play a role in hydrogen isotope partitioning (Graham et al. 1984). The issue of compositional effects on D/H fractionations remains largely unresolved and awaits further experimental studies. Solution composition The presence of dissolved species in the fluid phase can impact fractionation factors to a comparable or larger degree than the largest mineral composition effects. This effect applies specifically to fractionations between an aqueous fluid, and some other mineral,

Equilibrium Isotopic Fractionation Factors

25

gas or fluid phase. The seminal work on the isotopic solution effect, or ‘salt effect’ as it is commonly known, was done by H. Taube in the 1950s. This and subsequent work clearly demonstrated that the effects of many dissolved salts of geochemical interest on isotopic fractionation are non-trivial at or near room temperature (Taube 1954; Sofer and Gat 1972 1975; Stewart and Friedman 1975; Bopp et al. 1977; O'Neil and Truesdell 1991; Kakiuchi 1994). However, there was little information on the salt effect at elevated temperatures, and the available data were controversial with respect to the temperature and concentration-dependency of the effect (e.g. Truesdell 1974). In an attempt to resolve the controversy, several investigators carried out studies of oxygen and hydrogen isotope salt effects at elevated temperature, using a variety of experimental techniques (Matsuhisa et al. 1979, quartz- and albite-water at 600 and 700°C; Graham and Sheppard 1980, epidote-water 250-550°C; Kendall et al. 1983, calcite-water at 275°C; Kazahaya 1986, liquid-vapor to 345°C; Zhang et al. 1989, quartz-water at 180-550°C; Zhang et al. 1994, cassiterite- and wolframite-water at 200-420°C; Poulson and Schoonen 1994, dissolved HCO3-water at 100-300°C; Driesner 1996, Driesner and Seward 2000, liquidvapor at 50-413°C and calcite-water at 350-500°C; Kakiuchi 2000, liquid-H2O vapor). Some of these studies described a complex dependence of the isotope effects on temperature and solution composition with large uncertainties (Graham and Sheppard 1980; Kazahaya 1986; and Poulson and Schoonen 1994). Other studies observed little or no effect of dissolved salts on oxygen isotope fractionation in mineral-water systems (Matsuhisa et al. 1979; Kendall et al. 1983; Zhang et al. 1989, 1994). In perhaps the most comprehensive set of studies, Horita et al. (1993a,b; 1994, 1995a,b; 1996, 1997) investigated the effect of a number of dissolved salts, particularly NaCl, on isotopic partitioning from room temperature to 500°C. Terminology. The magnitude of the salt effect is conventionally represented by: Γ=

α A -aqueous soln α A-pure water

(15)

where A is a reference phase with which both pure water and the salt-bearing solution are exchanged in separate experiments. Water vapor is commonly used as the reference phase at lower temperatures, whereas minerals are used at higher temperature. Although Γ is formally defined as an activity-composition ratio (see Horita et al. 1993a for details), in practice, it is a measure of how much the fractionation factor between phase A and an aqueous fluid changes as material is dissolved into the fluid. Single salt solutions. Horita et al. (1993a,b; 1995a) determined values of Γ in single salt solutions (NaCl, KCl, MgCl2, CaCl2, Na2SO4, MgSO4) by means of the H2O(v)H2O(l) equilibration method at temperatures from 50 to 350°C. Figure 9 shows representative results from the 100°C experiments. For hydrogen isotopes, ΓH is greater than one for all of the salt solutions studied, and increases with salt concentration. Magnitudes of the effects are in the order CaCl2 ≥ MgCl2 > MgSO4 > KCl ≈ NaCl > Na2SO4 at the same molality. ΓO, on the other hand, is slightly less than or very close to 1, except for KCl solutions at 50°C. The measured salt effect trends for both hydrogen and oxygen are linear with the molalities of the salt solutions, and either decrease with ° temperature, or are nearly constant over the temperature range 50-100 C. Salt solutions of divalent cations (Ca and Mg) exhibit much larger oxygen isotope effects than those of monovalent cations (Na and K). Magnitudes of the oxygen isotope effects in NaCl solutions, and of hydrogen isotope effects in Na2SO4 and MgSO4, increase slightly from 50 to 100°C. The systematic changes of Γ with temperature and molality permit fitting of data for each salt to simple equations. A summary of these equations is given in Table 4. These data indicate that the identity of the cation largely controls oxygen isotope salt

26

Chacko, Cole & Horita

Figure 9. Experimental determined isotope salt effects (103 lnΓ) reported by Horita et al. (1993a) for (A) D/H and (B) 18O/16O fractionation at 100°C plotted against molality of the salt solution. The data were obtained by measuring the isotopic composition of water vapor over pure water and over salt solutions of the same isotopic composition. The solid lines are linear regressions with zero intercept through the data of each salt composition. Note that MgCl2 and CaCl2 have the largest (positive) D/H salt effects whereas MgCl2 and MgSO4 have the largest (negative) 18O/16O salt effect.

Table 4. Isotope salt effects determined by vapor-liquid equilibration. T Range (°C)

Salt

Isotopes

Function (m; molality & T: K)

NaCl:

18

D/H O/16O

103lnΓ=m(0.01680T-13.79+3255/T) 103lnΓ=m(-0.033+8.93x10-7T2-2.12x10-9T3)

10-350 10-350

KCl:

D/H 18 O/16O

103lnΓ=m(-5.1+2278.4/T) 103lnΓ=m(-0.612+230.83/T)

20-100 25-100

MgCl2:

D/H 18 O/16O

103lnΓ=4.14m 103lnΓ=m(0.841-582.73/T)

50-100 25-100

CaCl2:

D/H 18 O/16O

103lnΓ=m(0.0412T-31.38+7416.8/T) 103lnΓ=m(0.2447-211.09/T)

50-200 50-200

Na2SO4:

18

D/H O/16O

103lnΓ=0.86m 103lnΓ=-0.143m

50-100 50-100

MgSO4:

18

D/H O/16O

103lnΓ=m(8.45-2221.8/T) 103lnΓ=m(0.414-432.33/T)

50-100 0-100

Regressions are based on data reported by Horita et al. (1993a,b, 1995a). For the definition of Γ, see Equation (15).

Equilibrium Isotopic Fractionation Factors

27

effects in water. This can be rationalized by noting that cations interact strongly with the negatively-charged dipole of water molecules. Sofer and Gat (1972) pointed out that the ionic charge to radius ratio (ionic potential) of cations correlates positively with the measured oxygen isotope salt effects. For the same reason, it is expected that anions control hydrogen isotope salt effects, but their relationship is more complex. For example, there is a positive correlation between the radius of alkali and halogen ions and the magnitude of hydrogen isotope salt effects (Horita et al. 1993a). Complex salt solutions. Isotope salt effects have also been investigated in complex salt solutions consisting of mixtures of two or more salts in the system Na-K-Mg-Ca-ClSO4-H2O (Horita et al. 1993b). Some of the mixed salt solutions examined in that study were similar to natural brine compositions, such as from the Salton Sea geothermal system. The measured oxygen and hydrogen isotope salt effects in mixed salt solutions to very high ionic strengths (2-9) agree closely with calculations based on the assumption of a simple additive property of the isotope salt effects of individual salts in the solutions: 103 ln Γmixed salt soln = Σ (103 ln Γsingle salt soln) = Σ {mi(ai + bi/T)}

Figure 10. Calculated salt effects on D/H fractionation in the system kaolinite-water from 0 to 150°C (modified after Horita et al. 1993b). Calculations are based on the empirical equations given in Table 4. The dashed lines are the calculated curves for 1 and 3 molal NaCl solutions, and a synthetic Salton Sea brine (3.003 molal NaCl + 0.502 molal KCl + 0.990 molal CaCl2). The kaolinite-pure water curve (solid line) is from Liu and Epstein (1984). Note that neglect of the salt effect could result in large errors in calculated temperatures and/or fluid compositions when mineral-water fractionation data are applied to natural samples

Figure 10 illustrates the effect of salinity and salt composition on the kaolinite-water D/H fractionation factor, calculated assuming that the salt effect on minerals is the same as the salt effect on water vapor coexisting with the brine (Horita et al. 1993b). The effects of 1 and 3 m NaCl solutions and a representative Salton Sea brine composition (Williams and McKibben 1989) are shown as examples. The fractionation factor between kaolinite and pure water is taken from Liu and Epstein (1984). Note that the temperature of formation of a diagenetic phase calculated from the isotopic compositions of coexisting brine and mineral could be incorrect by as much as 80°C, or the calculated hydrogen isotope composition of the brine at a known temperature could be in error by as much as 15‰, if the salt effect on isotopic partitioning is ignored. Discrepancies reported in the literature between temperatures obtained from mineral-water isotope geothermometers, and temperatures derived from bore hole measurements, fluid inclusions, or other chemical geothermometers could result from neglect of isotope salt effects.

28

Chacko, Cole & Horita

Figure 11. (A) Hydrogen isotope salt effects in NaCl solutions to 450°C. Plotted points represent data obtained from liquid-vapor equilibration, H2-water equilibration, and mineral-water exchange experiments. (B) Oxygen isotope salt effects in NaCl solutions to 400°C. Plotted points represent data from liquid-vapor equilibration, CO2-water equilibration, and mineral-water exchange experiments. Solid curves are based on the liquid-vapor equilibration experiments of Horita et al. (1995a).

Equilibrium Isotopic Fractionation Factors

29

Salt effect at high temperature. Horita et al. (1995a), Berndt et al. (1996), Shmulovich et al. (1999) and Driesner and Seward (2000) extended the results for liquidvapor equilibration of NaCl solutions to 600°C. The results from Horita et al. (1995a) to 350°C are shown in Figure 11. The value of 103 ln Γ(18O/16O) is always negative and its magnitude increases with increasing temperature. In contrast, the magnitude of 103 ln Γ(D/H) decreases from 10°C to about 150°C, and then increases gradually to 350°C. The fractionation factors for both oxygen and hydrogen isotopes approach zero smoothly at the critical temperature of a given NaCl solution. The systematic nature of these results indicates that the complex temperature and concentration-dependencies of salt effects reported by Truesdell (1974) and Kazahaya (1986) are an experimental artifact. Another aspect to be considered in liquid-vapor experiments is that the vapor pressure and density of water vapor in equilibrium with pure water compared to that in equilibrium with NaCl solutions are different at a given temperature. With increasing temperature, water vapor becomes more dense and non-ideal, and the formation of water clusters (dimer, trimer, etc.) becomes significant. On the basis of molecular dynamics and ab initio calculations, Driesner (1997) suggested that water clusters partition hydrogen isotopes much differently than the water monomer. Thus, the effect of NaCl on liquidvapor isotope partitioning may reflect not only the isotope salt effect in liquid water, but also changes in the isotopic properties of water vapor. Furthermore, with increasing temperature, the concentration of NaCl in water vapor also increases, possibly causing an isotope salt effect in the vapor phase as well. Therefore, although the liquid-vapor equilibration method can provide the most precise results on the isotope salt effect, application of high-temperature data obtained with this technique may not be straightforward. In an attempt to document the salt effect on D/H fractionation at elevated temperatures, Horita et al. (unpublished) conducted a study of the system brucite (Mg[OH]2)-H2O±NaCl±MgCl2 from 200 to 500°C and salt concentrations up to 5 molal. Dissolved NaCl and MgCl2 consistently increased D/H fractionation factors by a small amount at all temperatures studied (Fig. 12). These data can be compared to data obtained on the epidote-water+NaCl system by Graham and Sheppard (1980). The results of the two studies are generally consistent, although the latter show slightly larger D/H effects. In contrast to the results for hydrogen isotopes, recent studies in the system calcite-H2ONaCl (Horita et al. 1995a; Hu and Clayton, in press) indicate that the oxygen isotope salt effect is negligible from 300-700°C and 1 to 15 kbar. Combined effects. Because temperature, pressure and solution composition all affect the physical properties of water, these three variables can act in concert to influence fractionation factors (Horita et al., in press; Hu and Clayton, in press). Increases in pressure and NaCl concentration both work to decrease the fluid’s affinity for deuterium. The isotopic effect is most pronounced at low pressures and NaCl concentrations, which relates to the fact that the largest changes in the density of the fluid occurs over this region of pressure-composition space. It appears that the fractionation factor and the density of aqueous NaCl solutions are closely related to each other. With additional systematic experiments, empirical equations can be designed that relate the isotope salt effects and the density of aqueous solutions. METHODS OF CALIBRATING FRACTIONATION FACTORS As noted at the beginning of the chapter, the four main methods for calibrating fractionation factors are theoretical calculations, semi-empirical bond-strength models, natural sample data, and laboratory experiments. Theoretical methods have already been

30

Chacko, Cole & Horita

Figure 12. The effect of NaCl on the hydrogen isotope fractionation (103 lnΓ) obtained from experiments involving liquid-vapor equilibration (solid bold curves) compared to those obtained from brucite-water (solid curves, Horita et al., in press) and epidote-water (dashed curves, Graham and Sheppard 1980) partial exchange results. Data from Horita et al. (in press); Solid circles = 1 molal NaCl, Solid squares = 3 molal NaCl, Solid triangles = 5 molal NaCl. Data from Graham and Sheppard (1980): Open, inverted triangles = 4 molal NaCl, Open circles = 1 molal NaCl.

discussed at length. Below, we critically summarize key elements of the other three calibration methods. Semi-empirical bond-strength calibration This method of calibration has been applied specifically to determining oxygen isotope fractionation factors involving minerals. The method is based on the observation 18 that the sequence of O enrichment in minerals is correlated with average cation-oxygen bond strengths in those minerals (Taylor and Epstein 1962; Garlick 1966). As noted previously, statistical mechanical theory predicts such a correlation because bond strength relates to vibrational frequency, and in turn ZPE differences between isotopic species. Bond-strength methods, which include the original methodology of Taylor and Epstein (1962), the site potential method (Smyth and Clayton 1988; Smyth 1989), and the increment method (Schütze 1980; Richter and Hoernes 1988; Hoffbauer et al. 1994; Zheng 1999a and references cited therein), attempt to quantify the relationship between bond strength and isotopic fractionation. All of these methods involve two major steps: (1) formulate an internally consistent measure of bond strength, and (2) link bond strength to fractionation factors. Taylor and Epstein method. Taylor and Epstein (1962) focussed on three major bond types in silicate minerals, the Si-O-Si bond (e.g. quartz), the Si-O-Al bond (e.g. anorthite) and the Si-O-M2+ bond (e.g. olivine). They assigned δ18O values to each of these bond types on the basis of their isotopic analyses of quartz, anorthite and olivine in igneous rocks, and suggested that the δ18O values of other silicates could be estimated through linear combination of these three bond types. Quartz-albite and quartz-diopside fractionation factors calculated with this method are in excellent agreement with those

Equilibrium Isotopic Fractionation Factors

31

determined experimentally (Clayton et al. 1989; Chiba et al. 1989). Savin and Lee (1988) and Saccocia et al. (1998) applied a modified version of this method to calculating fractionation factors for phyllosilicate minerals. Site-potential method. Smyth (1989) developed the site potential model for calculating electrostatic site energies associated with various anion sites in minerals. In essence, an anion’s site potential is the energy (in electron volts) required to remove that anion from its position in the crystal. Thus, an oxygen site potential provides a convenient monitor of how strongly bound that oxygen atom is within the mineral structure. Given data on individual oxygen sites in a mineral, a mean oxygen site potential for the mineral can be calculated from a weighted average of all of the oxygen sites. Smyth used this approach to calculate mean oxygen site potentials for 165 minerals. Following Smyth and Clayton (1988), Figure 13a plots experimentally determined quartz-mineral fractionation factors versus the difference in mean quartz-mineral site potentials. With the exception of forsterite, anhydrous and hydrous silicate minerals show a good linear correlation on this plot, which suggests that the trend may be useful in predicting fractionation factors for other silicates. Calcite, apatite and oxide minerals, however, fall well off the silicate trend. The poor fit of the oxide minerals can be attributed to neglect of cation mass in the site potential model (Smyth and Clayton 1988). It is well known that cation mass affects vibrational frequencies and frequency shifts on isotopic substitution. Thus, the significantly greater mass of cations in the oxide minerals relative to those in silicates would be expected to cause differences in their isotopic fractionation behavior. The deviation of calcite and apatite from the silicate trend is likely

Figure 13. Comparison of quartz-mineral fractionation factors given by bond-strength methods and experiments at 1000 K (modified after Chacko et al. 1996). (A) Comparison with the oxygen site potential model of Smyth (1989) where Vqtz and Vmineral are the mean oxygen site potentials of quartz and the mineral of interest, respectively. A least-squares regression through the origin fitted to all silicate data points except forsterite yields the equation: Δ1000K(qtz-mineral) = 0.751 (Vqtz – Vmineral). (B) Comparison of experimental data with the increment method calculations of Zheng (1993 1996 1997 1999a). Note that Zheng applied a low-temperature ‘correction’ factor (D) in his earlier papers (e.g., Zheng 1991 1993). That correction factor was not applied in some later studies (Zheng 1997 1999a). For internal consistency, the correction factor must be applied to all minerals or not at all. Following Zheng (1999a), the correction factor was not included in calculating fractionation factors for the minerals shown on the plot. The line represents 1:1 correspondence. Designations: Ab = albite; An = anorthite; Ap = apatite; Cal = calcite; Di = diopside; Fo = forsterite; Gh = gehlenite; Gr = grossular; Mt = magnetite; Mu = muscovite; Prv = CaTiO3-perovskite; Ru = rutile. Sources of experimental data: Clayton et al. (1989), Chiba et al. (1989), Gautason et al. (1993), Fortier and Lüttge (1995), Rosenbaum and Mattey (1995), and Chacko et al. (1989, 1996).

32

Chacko, Cole & Horita

fractionation behavior. The deviation of calcite and apatite from the silicate trend is likely due to the strongly covalent nature of bonding in carbonate and phosphate functional groups (Smyth and Clayton 1988). The site potential model, on the other hand, assumes that bonding in crystals is fully ionic (Smyth 1989). The site potential method also provides no direct indication of the temperature-dependence of isotopic fractionation. Increment method. The increment method is also based on bond strengths but attempts to incorporate the effect of cation mass in its parameterization. Although they differ in detail, the various formulations of this method assign 18O increment values (ict-O) for individual cation-oxygen bonds, which are then referenced to the increment value for the silicon-oxygen bond. Increment values are calculated from data on cation valence and coordination number, and cation-oxygen bond lengths. Cation mass effects are incorporated by treating the cation-oxygen pair as a diatomic molecule, and calculating the change in the reduced mass of this molecule on 18O substitution (Eqn. 4). Empirically derived parameters are also included for strongly-bonded and weakly-bonded cations, and the OH- group in hydrous silicates is treated separately. The total 18O increment value for a mineral (I-18O), which is a weighted average of the increment values of its constituent cation-oxygen bonds, is an indication of the mineral’s relative affinity for 18 O. To calculate fractionation factors, this relative scale of 18O enrichment must be linked to an absolute scale derived from experiments or statistical mechanical theory. The reduced partition ratios of quartz obtained from experiments or theory have generally been used to make this link. Figure 13b shows a comparison of fractionation factors at 1000 K given by the increment method calculations of Zheng (1993 1996 1997 1999a) with those indicated by experiments. There is good agreement between the two approaches for anhydrous silicates. However, hydrous silicates, apatite, and to a lesser extent, oxide minerals, deviate from the 1:1 correspondence line. Comparisons with other formulations of the increment method (e.g. Hoffbauer et al. 1994) give broadly similar results. These comparisons suggest that, although useful for anhydrous silicates, the increment method in its current form does not adequately deal with the effects of cation mass, hydroxyl groups or covalent bonding on fractionation behavior. Summary. Bond-strength methods are in wide use because of the relative ease of determining fractionation factors for a large number of minerals with this approach. It must be emphasized, however, that these methods do not comprise a separate theory of isotopic fractionation. They are a derivative approach in which bond strengths serve as a proxy for the vibrational energies that are the root cause of fractionation. Thus, at best, bond-strength methods are only as good as the various parameterizations that go into linking bond strengths to vibrational energies, and ultimately to fractionation factors. In general, the mathematical forms of these parameterizations are not rigorously grounded in theory. The statements above are not meant to imply that bond-strength methods have no value. In cases where experiments have not been done or where theoretical calculations have not been made, these methods may provide a reasonable interim estimate of fractionation factors. Such estimates must, however, be regarded with caution until confirmed by independent methods. Natural sample calibration

In principle, isotopic analyses of natural samples can also be used to calibrate fractionation factors. Effective use of the natural sample method, however, requires that several stringent criteria be met. (1) The phases being calibrated first attained isotopic equilibrium at some temperature, and subsequently retained their equilibrium isotopic compositions, (2) equilibration temperatures in the samples of interest are well determined by independent methods, and (3) the geothermometers used to determine

Equilibrium Isotopic Fractionation Factors

33

temperature equilibrated (or re-equilibrated) at the same conditions as the isotopic system being calibrated. Rigorous application of these criteria significantly limits the number of samples suitable for use with this calibration method (cf. Kohn and Valley 1998b,c). Moreover, for many samples, it is difficult to demonstrate unambiguously that these criteria have been met. Despite the potential pitfalls, the natural sample method has been widely applied, and can in favorable cases provide important insights on isotope fractionation behavior. In this regard, Valley (this volume) describes how microanalytical techniques can be used to select the optimal samples for use in natural sample calibrations. Probably the most widely used set of natural sample calibrations is that of Bottinga and Javoy (1975). With oxygen isotope data on minerals from a large number of igneous and metamorphic rocks serving as input, these workers followed a bootstrap procedure for calibrating fractionation factors. More specifically, they estimated temperatures for individual samples with modified laboratory calibrations (Bottinga and Javoy 1973) of oxygen isotope geothermometers comprising feldspar, and one of quartz, muscovite or magnetite. They then empirically calibrated fractionation factors for other minerals (pyroxene, olivine, garnet, amphibole, biotite, and ilmenite) present in the same samples. The validity of this approach depends critically on whether all the minerals involved in the calibration preserved their equilibrium isotopic compositions (Clayton 1981). On the basis primarily of concordancy of the isotopic temperatures that they obtained, Bottinga and Javoy (1975) argued that most of the samples that they examined were in fact in isotopic equilibrium. Deines (1977) came to the opposite conclusion upon detailed statistical evaluation of the same body of isotopic data. It is also now well established from laboratory diffusion data, and from numerical modeling of natural sample data that the minerals involved in these calibrations have markedly different susceptibilities to reequilibration, and thus are not likely to be in equilibrium in samples that have cooled slowly from high temperature (e.g. Giletti 1986; Eiler et al. 1992; Farquhar et al. 1993; Jenkin et al. 1994; Kohn and Valley 1998b,c). Therefore, purely on theoretical grounds, the approach to natural sample calibration taken by Bottinga and Javoy (1975) is suspect. Nevertheless, some of the mineral-pair fractionation factors reported in that study are in good agreement with the best available experimental calibrations. This agreement is probably fortuitous as two out of the three reference calibrations (feldspar-muscovite, feldspar-magnetite) used in the Bottinga and Javoy study are in substantial disagreement with the same set of experimental data. The natural sample method is most likely to be successful with rocks formed at low temperatures, with rapidly cooled volcanic rocks, and with isotopically refractory minerals (e.g. garnet, graphite) in more slowly cooled rocks. Minerals formed in lowtemperature environments are less susceptible to diffusional re-equilibration on cooling. They can, however, undergo changes in their isotopic compositions through recrystallization during processes such as diagenesis or deformation. Provided that this has not occurred, and that the mineral or minerals being calibrated have not become isotopically zoned during growth, the fractionations measured in such samples may yield reliable estimates of equilibrium fractionation factors. This approach to calibration also requires that the initial formation conditions of the mineral (temperature, the isotopic composition of the fluid, etc.) are well characterized. Examples of natural calibrations using low-temperature samples include the calcite-H2O and gibbsite-H2O oxygen isotope calibrations of Epstein et al. (1953) and Lawrence and Taylor (1971), respectively. The very rapid cooling associated with the formation of volcanic rocks makes it likely that any phenocryst minerals present in these rocks retain their original isotopic composition. Thus, if these phenocrysts initially crystallized in isotopic equilibrium (cf.

34

Chacko, Cole & Horita

well suited for use in natural sample calibrations. We are not aware of any calibrations based exclusively on the analysis of minerals in volcanic rocks. In part because of improvements in analytical techniques over the past decade, there has been an increase in natural sample calibrations involving refractory minerals found in slowly cooled metamorphic and igneous rocks. Examples include calcite-graphite carbon isotope fractionations (Valley and O’Neil 1981; Wada and Suzuki 1983; Dunn and Valley 1992; Kitchen and Valley 1995), and garnet-zircon, garnet-pyroxene, garnetstaurolite, garnet-kyanite, and quartz-kyanite oxygen isotope fractionations (Valley et al. 1994; Sharp 1995; Kohn and Valley 1998b,c). The low diffusion rates of oxygen or carbon in these minerals greatly reduce the possibility of re-equilibration effects on cooling, and in that respect make them well suited for natural sample calibration. A concern with highly refractory mineral pairs, however, is the persistence of lowertemperature isotopic compositions during prograde metamorphic evolution. A possible example of this problem is found in the calcite-graphite system. The temperature coefficient of fractionations in this system given by two independent theoretical calibrations (Chacko et al. 1991; Polyakov and Kharlashina 1995) is in excellent agreement with that given by two high-temperature (T > 650°C) natural sample calibrations (Valley and O’Neil 1981; Kitchen and Valley 1995) (Fig. 14). In contrast, three other natural sample calibrations (Wada and Suzuki 1983; Morikiyo 1984; Dunn and Valley 1992), based primarily on data from lower-temperature samples (T = 270650°C), indicate temperature coefficients that are higher by a factor of 1.5-2.0. Extrapolation of the lower temperature natural sample calibrations to the hightemperature limit (ln α = 0) requires the shape of the calcite-graphite fractionation curve

Figure 14. Comparison of theoretical (Chacko et al. 1991; Polyakov and Kharlashina 1995) and natural sample (Valley and O’Neil 1981; Wada and Suzuki 1983; Morikiyo 1984; Dunn and Valley 1992; Kitchen and Valley 1995) calibrations of the calcite-graphite fractionation factor. Note that the three lower-temperature natural sample calibrations (WS, M and DV) require a much different shape for the fractionation curve than indicated by the theoretical calculations. See text for discussion.

Equilibrium Isotopic Fractionation Factors

35

to be distinctly convex towards the temperature axis. This is in marked contrast to the weakly concave shape indicated by the theoretical calculations. Although the approximations required in making theoretical calculations result in significant uncertainty in the absolute magnitude of calculated fractionations, the calculations do place strict constraints on the basic shape of fractionation curves (Clayton and Kieffer 1991; Chacko et al. 1991). It is unlikely, therefore, that the shape of the fractionation curve implied by the Wada and Suzuki (1983), Morikiyo (1984), and Dunn and Valley (1992) calibrations is correct. Chacko et al. (1991) suggested that the deviation of the low-temperature natural sample data points from the calculated curve is due to the incomplete equilibration of calcite-graphite pairs at temperatures below about 650°C (see also discussion in Dunn and Valley 1992). If this interpretation is correct, calcite-graphite pairs formed at lower temperatures are unsuitable for use in natural sample calibrations, or for stable isotope thermometry. Similar problems can occur in refractory silicate mineral pairs if the two minerals being calibrated formed at different temperatures, or if either one of the minerals formed over a range of temperatures in a prograde metamorphic sequence. In the first case, isotopic equilibrium may never have been established between the two minerals, whereas, in the second case, one or both minerals may be isotopically zoned. Sharp (1995), Kohn and Valley (1998c) and Tennie et al. (1998) noted these possibilities for natural sample calibrations involving the highly refractory minerals kyanite and garnet. Kohn and Valley (1998c) argued convincingly, however, that such problems can be overcome by carefully selecting samples with the appropriate textural and petrological characteristics. This screening process requires a detailed understanding of the metamorphic reaction history of samples. Experimental calibration Philosophy and methodology of experiments. Laboratory experiments are the most direct method of calibration in that they require the least number of a priori assumptions, and also generally permit control of the variables that may influence fractionation factors. The other calibration methods, although perhaps valid, must be regarded as tentative until confirmed by laboratory experiments. The reader is referred to the detailed reviews of experimental methods provided by O’Neil (1986) and Chacko (1993) as only the major points are summarized here.

A useful frame of reference for discussing experimental methodology is to consider the design of an ideal experiment. Such an experiment would involve direct isotopic exchange between the two substances of interest. For example, the ideal experiment for determining the oxygen isotope fractionation factor between olivine and orthopyroxene would be one in which the two minerals are intimately mixed and allowed to equilibrate at the desired temperature until isotopic equilibrium is established. To confirm the attainment of equilibrium, a companion experiment would be carried out at the same temperature consisting of olivine and orthopyroxene with the same chemical composition and structural state but with an initial isotopic fractionation on the opposite side of the equilibrium value. Obtaining the same olivine-orthopyroxene fractionation factor in both experiments would constitute a successful experimental reversal. Additional criteria for an ideal experiment include no chemical and textural changes in the starting materials during the course of the experiment. That is, isotopic exchange is accomplished exclusively through a diffusional process, which in turn is driven solely by the free energy change for the exchange reaction. For practical reasons, most, if not all, experimental studies depart to some extent from this ideal experiment. Firstly, a direct exchange between the phases of interest is

36

Chacko, Cole & Horita

often not possible. For example, it would be impossible to obtain a clean physical separation between olivine and orthopyroxene (for isotopic analysis) after these minerals had been ground to a sufficiently fine grain size to yield reasonable amounts of isotopic exchange on laboratory time-scales. As a result, fractionation factors are often determined indirectly by exchanging phases with a common isotopic exchange medium in separate experiments, and combining the resulting data to obtain the fractionation factor of interest. The exchange media that have been used to date, H2O, CO2, H2, CaCO3, and BaCO3 were chosen partly because they are easily separated from the phase of interest by physical or chemical techniques after the exchange experiment. A second problem that plagues many exchange experiments is the sluggishness of exchange rates. In particular, the diffusion rates of oxygen and carbon in many minerals are slow enough to preclude a close approach to isotopic equilibrium in typical laboratory time scales through purely diffusional processes. As a consequence, many experiments in oxygen and carbon isotope systems induce exchange either by recrystallization of preexisting phases or crystallization of new phases (synthesis) during the experiment. Both procedures result in an experiment that is less than ideal. Synthesis experiments generally involve the inversion of a polymorph or crystallization of gels or oxide mixes. The free energy change associated with these processes is about 1000 times greater than those associated with isotope exchange reactions (Matthews et al. 1983b). These processes can also be very rapid and unidirectional, possibly resulting in kinetic rather than equilibrium isotopic fractionations. This was demonstrated elegantly by Matsuhisa et al. (1978) in hydrothermal experiments at 250°C, where quartz-water fractionations obtained by inverting cristobalite to quartz were about 3‰ different than those obtained by direct quartz-water exchange experiments. They interpreted these results to reflect a kinetic isotope effect associated with the crystallization of quartz from cristobalite. Similar non-equilibrium effects were noted by Matthews et al. (1983a) in crystallizing wollastonite and diopside from mixes of constituent oxides. In other cases, synthesis and direct exchange experiments appear to give comparable results (O’Neil and Taylor 1967; Matsuhisa et al. 1979; Lichtenstein and Hoernes 1992; Scheele and Hoefs 1992). Despite the problems inherent to the mineral synthesis technique, this technique may be the only viable means of obtaining reasonable amounts of isotopic exchange in low temperature experiments (T < 250°C). With care, the technique may indeed provide reliable equilibrium fractionation factors. Criteria that have been used to suggest an approach to equilibrium fractionations in such experiments include controlled precipitation rates, an understanding of reaction pathways and the likelihood of kinetic fractionations associated with particular pathways, similar fractionations obtained in synthesis from different starting materials, and a concurrence of fractionation factors obtained with synthesis and direct exchange techniques (e.g. O’Neil and Taylor 1969; Scheele and Hoefs 1992; Bird et al. 1993; Kim and O’Neil 1997; Bao and Koch 1999). It should be emphasized, however, that even when these criteria are met, synthesis experiments cannot rigorously demonstrate the attainment of isotopic equilibrium. Recrystallization of existing phases during the course of an experiment also involves driving forces other than solely those of the exchange reaction. However, the free energy change that drives recrystallization is approximately the same order of magnitude as that associated with isotopic exchange (Matthews et al. 1983b). As such, experiments involving recrystallization are less likely to incorporate kinetic effects than those requiring the growth of new phases. In contrast to this view, Sharp and Kirschner (1994) argued that recrystallization based experiments involving quartz do incorporate significant kinetic effects. In general, however, recrystallization experiments, although less than ideal, are clearly preferable to synthesis experiments, and often provide the best

Equilibrium Isotopic Fractionation Factors

37

available means of calibrating fractionation factors at moderate to high temperature. Two additional features of isotope exchange experiments are useful in the acquisition of equilibrium fractionation data. Firstly, there is a substantial increase in oxygen isotope exchange rates with increasing pressure (Clayton et al. 1975; Matthews et al. 1983a,b; Goldsmith 1991). Thus, for a given experimental run time, experiments performed at higher pressure show a closer approach to equilibrium and thereby provide more tightly constrained data. Secondly, because isotope exchange reactions run in both forward and reverse directions involve exchange of the same element between phases, it is reasonable to assume that they have nearly identical reaction rates. On the basis of this assumption, Northrop and Clayton (1966) developed the following equation for extracting equilibrium oxygen and carbon isotope fractionation factors from partially exchanged experimental data: ln αi = ln αeq - 1/F (ln αf - ln αi) where the superscripts i, f and eq refer to the initial, final and equilibrium fractionations, respectively, and F is the fractional approach to equilibrium during the experiment. This equation indicates that if forward and reverse directions of an isotope exchange reaction have the same exchange rates, then data from three or more companion experiments run at the same conditions, and for the same amount of time should be linear when plotted as ln αi versus (ln αf - ln αi). The slope of this line gives the fractionation approach to equilibrium and the y intercept gives the extrapolated equilibrium fractionation factor. This method becomes progressively more reliable as F approaches unity because the intercept between the plotted line and the y-axis occurs at higher angles. The slightly modified version of this equation developed for hydrogen isotope exchange experiments is (Suzuoki and Epstein 1976): (αi - 1) = (αeq - 1) - 1/F (αf - αi) More recently, Criss (1999, p. 204) derived a different equation for extracting an equilibrium fractionation factor (αA-B) from a pair of exchange experiments (labeled 1 and 2) that has not proceeded completely to equilibrium: (1 − H)(α eq )2 + [Hα 1i + Hα 2f − α 2i − α 1f ](α eq ) + [α 1f α 2i − Hα 2f α1i ] = 0 where  R iB   R Bf     H=  R Bf  1  R iB  2 and RB is the isotope ratio (e.g. 18O/16O) of phase B. The quadratic formula is used to solve the equation for αeq. Criss (1999) proposed that this equation yields more exact and more correct results than the commonly used equation of Northrop and Clayton (1966). In practice, the difference in equilibrium fractionation factors obtained with the two equations is small in most cases. Hydrothermal- and carbonate-exchange techniques. The majority of available experimental fractionation data are for oxygen isotope fractionations involving minerals. Much of the data, particularly the early data, were obtained using water as the isotopic exchange medium. These experiments were either done at ambient pressure (typically synthesis experiments), or in cold-seal pressure vessels at pressures of 1 to 3 kbar (e.g. O’Neil and Taylor 1967; O’Neil et al. 1969; Clayton et al. 1972). Later experiments were done in a piston cylinder apparatus (at ~15 kbar) to exploit the pressure enhancement of exchange rates (Clayton et al. 1975; Matsuhisa et al. 1979; Matthews et al. 1983a,b).

38

Chacko, Cole & Horita

Even though there were some differences in experimental methodologies, the agreement between the various studies was in general excellent for some of the major mineral-water systems (e.g. albite-water, quartz- and calcite-water systems). Matthews et al. (1983a) and Matthews (1994) provide compilations of mineral-pair fractionation factors derived from the high-temperature hydrothermal experiments. Despite the apparent success of the hydrothermal technique, there are a number of theoretical and practical drawbacks to this method of experimentation. Under hydrothermal conditions, some minerals dissolve excessively, melt, or react to form hydrous phases, making them unsuited to this type of investigation. Furthermore, the high vibrational frequencies of the water molecule results in complex temperaturedependencies for mineral-H2O fractionations (Fig. 3), thereby making it difficult to extrapolate fractionations outside the experimentally investigated temperature range. To overcome some of these difficulties, Clayton et al. (1989) developed a technique that uses CaCO3 rather than water as the common isotope exchange medium. Rosenbaum et al. (1994) used a variation of this technique with BaCO3 as the exchange medium. Advantages of the so-called carbonate-exchange technique include: (1) the ability to carry out experiments at high temperatures (up to 1400°C) because of the high thermal stability of many mineral-carbonate systems, (2) the avoidance of problems associated with mineral solubility and its potential effect on fractionation factors, (3) ease of mineral separation, and (4) the relative ease and high precision of carbonate isotopic analysis. Extrapolation of experimental data is also simplified in that anhydrous mineral-carbonate fractionations should be approximately linear through the origin on fractionation plots at temperatures above 600°C (Clayton et al. 1989). There now exists a large body of oxygen isotope fractionation data for minerals acquired with the carbonate-exchange technique (summarized in Table 5). Other experimental techniques. There have been several other recent advances in experimental techniques for determining fractionation factors for oxygen, carbon, and hydrogen isotope systems. CO2 has successfully been used as an exchange medium for determining carbon or oxygen isotope fractionation factors for melts, glasses and carbonate, silicate and oxide minerals at both high and low pressure (Mattey et al. 1990; Chacko et al. 1991; Stolper and Epstein 1991; Scheele and Hoefs 1992; Matthews et al. 1994; Rosenbaum 1994; Palin et al. 1996; Fayek and Kyser 2000). Vennemann and O’Neil (1996) used H2 gas as an exchange medium in low-pressure experiments to obtain hydrogen isotope fractionation factors for hydrous minerals. Horita (2001) was able to obtain tightly reversed carbon isotope fractionation data in the notoriously sluggish CO2CH4 system by using transition-metal catalysts to accelerate exchange rates. Fortier et al. (1995) and Chacko et al. (1999) used the ion microprobe to obtain oxygen and hydrogen isotope analyses, respectively, of run products in magnetite-water and epidote-water exchange experiments. The latter study was novel in that it used millimeter-sized, single crystals of epidote instead finely ground epidote powder as starting material, and then analyzed the outer 0.5-2 μm of the isotopically exchanged crystals with an ion microprobe. The advantage of such an approach is that it allows D/H fractionation factors to be determined in experiments in which most or all of the isotopic exchange occurs by a diffusional process, rather than by a combination of diffusion and recrystallization (i.e. an ideal exchange experiment). The disadvantage of the approach is that the precision of hydrogen isotope analyses on the ion microprobe is currently a factor of 2 to 4 poorer than can obtained by conventional methods. That precision will no doubt improve with advances in instrumentation. Figure 15 shows Chacko et al.’s (1999) data for the epidoteH2O system. Despite initial fractionations that were in most cases far from equilibrium, the companion experiments at each temperature were successful in bracketing the equilibrium fractionation factor within analytical error. This technique should be useful

Cal 0.38

Ab 0.94 0.56

Mu 1.37 0.99 0.43

F-Phl 1.64 1.26 0.70 0.27

An 1.99 1.61 1.05 0.62 0.35

Phl 2.16 1.78 1.22 0.79 0.52 0.17

*Ap 2.51 2.13 1.57 1.14 0.87 0.52 0.35

Di 2.75 2.37 1.81 1.38 1.11 0.76 0.59 0.24

Gr 3.15 2.77 2.21 1.78 1.51 1.72 0.99 0.64 0.40

Gh 3.50 3.12 2.56 2.13 1.86 1.51 1.34 0.99 0.75 0.35

Fo 3.67 3.29 2.73 2.30 2.03 1.68 1.51 1.16 0.92 0.52 0.17

Ru 4.69 4.31 3.75 3.32 3.05 2.70 2.53 2.18 1.94 1.54 1.19 1.02

Mt 6.29 5.91 5.35 4.92 4.65 4.30 4.13 3.78 3.54 3.14 2.79 2.62 1.60

Prv 6.80 6.42 5.86 5.43 5.16 4.81 4.64 4.29 4.05 3.65 3.30 3.13 2.11 0.51

*The equation for apatite is not the same as that given in Fortier and Lüttge (1995). Following Chacko et al.’s (1996) suggestion for hydrous minerals, Fortier and Lüttge’s 500-800°C apatite-calcite fractionation data have been regressed by a straight line through the origin on a plot of 1000 ln α vs. 106/T2.

Equations should not be extrapolated below ~600°C. All data derived from experiments using the carbonate-exchange technique. Qtz = quartz; Cal = calcite; Ab = albite; Mu = muscovite; F-Phl = fluorophlogopite; An = anorthite; Phl = hydroxyphlogopite; Ap = apatite; Di = diopside; Gr = grossular; Gh = gehlenite; Fo = forsterite; Ru = rutile; Mt = magnetite; Prv = CaTiO3 - perovskite. Data from Clayton et al. (1989), Chiba et al. (1989); Chacko et al. (1989, 1996), Gautason et al. (1993), Fortier and Lüttge (1995), and Rosenbaum and Mattey (1995).

Coefficients for mineral-pair fractionation factors of the form 1000 ln α = A x 106 /T2(K), where the coefficient A is given in the table.

Qtz Cal Ab Mu F-Phl An Phl Ap Di Gr Gh Fo Ru Mt

Table 5. Coefficients for mineral-pair oxygen isotope fractionation factors.

Equilibrium Isotopic Fractionation Factors 39

40

Chacko, Cole & Horita

Figure 15. Experimental D/H fractionation data in the epidotewater system plotted as a function of 106 T-2 (modified after Chacko et al. 1999). The experiments involved exchange between water and large, single crystals of epidote. The isotopically exchanged crystals were analyzed by ion microprobe. Note that equilibrium fractionation factor was bracketed within analytical error (±6-10‰) to temperatures as low as 300°C.

for determining D/H fractionation factors in other mineral-water systems. On a more general level, micro-analytical techniques such as the ion microprobe hold great promise for isotope exchange experiments in that they open up the possibility of extending these experiments to significantly lower temperature. SUMMARY OF FRACTIONATION FACTORS

Appendices 2-4 are an annotated list of published experimental and natural sample calibrations of oxygen, carbon and hydrogen isotope fractionation factors applicable to geological systems. The reader is referred to the Introduction section for sources of information on fractionation factors derived from theoretical and bond-strength methods. In this section, we extract from the larger tabulation some sets of fractionation data that have wide applicability. Our goal is to provide an overview of the current state of knowledge on these key fractionations, and to highlight areas where more work needs to be done. A summary of this type is necessarily subjective, and the reader is encouraged to consult the references given in this section and in the bibliography for alternative points of view. Oxygen isotope fractionation factors Mineral-pair fractionations. As detailed above, there currently exists two large sets of experimental data on oxygen isotope fractionation factors between minerals, one obtained using water, and the other using carbonate as the isotopic exchange medium. Mineral-pair fractionation factors derived from the two data sets are generally in good agreement except for fractionations involving quartz or calcite (Clayton et al. 1989; Chiba et al. 1989; Chacko 1993, Matthews 1994). The discrepancies are enigmatic because both hydrothermal and carbonate-exchange experiments meet many of the criteria listed above for an ideal exchange experiment. Moreover, the discrepancies cannot be attributed to inter-laboratory differences as most of both data sets were

Equilibrium Isotopic Fractionation Factors

41

acquired at the same laboratory. Hu and Clayton (in press) provided a resolution to this paradox with an experimental investigation of the oxygen isotope salt effect on quartzH2O and calcite-H2O fractionations at high pressure and temperature. They found a significant salt effect in the quartz-H2O system but no salt effect in otherwise comparable experiments in the calcite-H2O system. The implications of these results are profound. If salt was the only dissolved constituent in the aqueous fluid, the magnitude of the salt effect (Γ) should be identical in both sets of experiments. The fact that Γqtz-fluid is different than Γcalcite-fluid indicates that minerals also dissolve appreciably at experimental conditions, and importantly, that the nature of the dissolved mineral species significantly affects fractionations in mineral-H2O systems. Thus, mineral-pair fractionation factors derived from different sets of mineral-H2O experiments may not be internally consistent because the characteristics of the fluid phase may vary with the mineral under investigation. To ensure internal consistency, two minerals must be exchanged with the same fluid. Hu and Clayton (in press) tested this idea with three-phase experiments in the quartz-calcite-H2O and phlogopite-calcite-H2O systems. The mineral-fluid fractionation factors derived from these three-phase experiments are different than those obtained in two-phase mineral-fluid experiments. However, the quartz-calcite and phlogopite-calcite fractionation factors derived from the three-phase experiments are the same as those obtained in the two-phase mineral-calcite experiments of Clayton et al. (1989) and Chacko et al. (1996). These results indicate that, in a choice between the two data sets, the carbonate-exchange technique provides the better mineral-pair fractionation data.

Figure 16. Summary of mineral-calcite fractionation factors given by the carbonate-exchange technique. All data have been fit by straight lines through the origin. Abbreviations and sources of data given in Table 5.

Mineral-calcite fractionation factors obtained with the carbonate-exchange technique are shown in graphical form in Figure 16. Each fractionation line shown on the plot is constrained by sets of experiments conducted at two to five different temperatures from 500 to 1300°C. These mineral-calcite fractionation data can be combined to obtain a large matrix of mineral-pair fractionation factors (Table 5). The straight-line equations given in Table 5 are adequate for calculating fractionation factors at high temperature but

42

Chacko, Cole & Horita

should not be used at temperatures of below about 600°C. Clayton and Kieffer (1991) developed a methodology for extrapolating these fractionation data to lower temperatures. With the calculated partition function ratios for calcite serving as a baseline (Chacko et al. 1991), they used the theoretical calculations (Becker 1971; Kieffer 1982) to constrain the basic shape of each fractionation curve. They then applied a correction factor to the calculated partition function ratios of individual minerals to optimize agreement between theory and experiment. It should be noted that the magnitude of the correction factor required to bring the calculations into agreement with the experiments was small for all minerals except rutile (Clayton and Kieffer 1991; Chacko et al. 1996). The advantage of this fitting procedure is that it allows experimental data, which necessarily must be obtained over a limited temperature range, to be extrapolated to lower temperature in a manner that is theoretically justifiable. Table 6 gives polynomial expressions for calculating the reduced partition function ratios for minerals derived through this fitting procedure. Table 6. Reduced partition function ratios for minerals. Mineral

1000 ln β

Calcite

11.781 x - 0.420 x2 + 0.0158 x3 12.116 x - 0.370 x2 + 0.0123 x3

Quartz Albite Muscovite Anorthite Phlogopite F-phlogopite Diopside Forsterite Rutile

11.134 x - 0.326 x2 + 0.0104 x3 10.766 x - 0.412 x2 + 0.0209 x3 9.993 x - 0.271 x2 + 0.0082 x3 9.969 x - 0.382 x2 + 0.0194 x3 10.475 x - 0.401 x2 + 0.0203 x3 9.237 x - 0.199 x2 + 0.0053x3 8.236 x - 0.142 x2 + 0.0032 x3 7.258 x – 0.125 x2 + 0.0033 x3

5.674 x - 0.038 x2 + 0.0003 x3 Polynomial expressions describing oxygen reduced partition function ratios 6 2 (1000 ln β) for minerals where x= 10 /T (K). The equations are only applicable at temperatures above 400K. The fractionation factor between any two phases at a particular temperature is given by the algebraic difference in their 1000 lnβ values (1000 ln βA – 1000 ln βB). Expressions are from Clayton and Kieffer (1991) and Chacko et al. (1996). Magnetite

Mineral-pair fractionation factors obtained with the carbonate-exchange technique have gained wide, but not universal, acceptance. In particular, Sharp and Kirschner (1994) questioned the validity of quartz-calcite fractionations given by this technique. Their natural sample calibration of this important system gave systematically larger fractionations than indicated by the experiments of Clayton et al. (1989). They attributed the discrepancy to a kinetic isotope effect in the experiments engendered by the rapid rate of quartz recrystallization relative to the rate of oxygen diffusion in the calcite exchange medium. We note, however, that diffusion rates of oxygen in quartz and calcite at 1000°C (Giletti and Yund 1984; Farver 1994) are rapid enough to isotopically homogenize small grains (radii < 3 μm) of these minerals by volume diffusion in 24 hours. Thus, the 1000°C quartz-calcite experiments, which used fine grained starting materials (diameter

Equilibrium Isotopic Fractionation Factors

43

1 to 10 µm) and were held at temperature for 24 hours, were of sufficiently long duration to establish a true diffusional equilibrium between the two minerals. The fractionation factor obtained by Clayton et al. in these high-temperature experiments is entirely consistent with their results at lower temperature (600-800°C), and does not agree with that given by the natural sample calibration. This suggests that the difference between the two calibrations has some other fundamental cause than the one proposed by Sharp and Kirschner. We believe the bulk of the evidence favors the experimental calibration. Nevertheless, given its importance, it seems prudent to attempt to re-determine fractionation factors in the quartz-calcite system with an independent method. For example, it may be possible to indirectly determine quartz-calcite fractionations through a combination of data from CO2-quartz and CO2-calcite experiments at high pressure and temperature. Another mineral that bears further investigation is kyanite. Existing experimental and natural sample calibrations of quartz-kyanite oxygen isotope fractionations are widely discrepant (Sharp 1995; Tennie et al. 1998), as are three bond-strength estimates of these fractionations (Smyth and Clayton 1988; Hoffbauer et al. 1994; Zheng 1999a). Although the experimental data for kyanite were obtained with the carbonate-exchange technique (Tennie et al. 1998), these data are not entirely compatible with the rest of the carbonate-exchange data set because the experiments involved polymorphic inversion of andalusite to kyanite, rather than direct kyanite-carbonate exchange. As such, it cannot unambiguously be shown that the fractionations measured in the experiments represent true equilibrium values. The highly refractory nature of kyanite in terms of oxygen isotope exchange makes it a difficult mineral to work with from both an experimental and natural sample perspective. However, that same characteristic potentially makes it a very useful mineral for elucidating the metamorphic history of rocks. Additional studies, although difficult, should be pursued. Mineral-fluid oxygen isotope fractionations at low temperature. Unlike the case for high-temperature fractionations between minerals, the current status of our knowledge on oxygen isotope fractionation between minerals and fluids/gases at low-temperatures (10‰) variation especially at temperatures below 40°C, depending on reaction pathways and solution compositions (e.g. pH). Mineral precipitates at low temperatures tend to be

Equilibrium Isotopic Fractionation Factors

45

Figure 18. Plot of oxygen isotope fractionation factor between magnetite and water at low temperatures. Data sources: Becker and Clayton (1976) and Rowe et al. (1994) - empiricaltheoretical calculations; Zheng (1995) empirical increment method; O’Neil and Clayton (1964) - extrapolation of high-temperature experimental results to an analysis of magnetite teeth from marine chiton; Blattner et al. (1983) – natural samples from active geothermal system in New Zealand; Zhang et al. (1997) – extracellular magnetite produced by Fe(III)-reducing bacteria in culture; Mandernack et al. (1999) – magnetite produced intracellular by magnetotactic bacteria in culture.

Figure 19. Plot of oxygen isotope fractionation factor between Fe(III) oxides (hematite and goethite) and water at low temperatures. Data sources: Zheng (1991, 1998) – increment method; Clayton and Epstein (1961), and Clayton (1963) – empirical calibration based on natural samples; Yapp (1990), Müller (1995), and Bao and Koch (1999) – laboratory synthesis experiments via different reaction pathways.

poorly crystalline and fine-grained (400°C. Geochim Cosmochim Acta 64:1773-1784 Dugan JP, Borthwick J (1986) Carbon dioxide-water oxygen isotope fractionation factor using chlorine trifluoride and guanidine hydrochloride techniques. Analyt Chem 58:3052-3054 Dugan JP, Borthwick J, Harmon RS, Gagnier MA, Glahn JE, Kinsel EP, MacLeod S, Viglino JA (1985) Guanidine hydrochloride method for determination of water oxygen isotope ratios and the oxygen-18 fractionation between carbon dioxide and water at 25°C. Analyt Chem 57:1734-1736 Dunn SR, Valley JW (1992) Calcite-graphite isotope thermometry: a test for polymetamorphism in marble, Tudor gabbro aureole, Ontario, Canada. J Metam Geol 10:487-501 Eiler JM, Baumgartner LP, Valley JW (1992) Intercrystalline stable isotope diffusion: a fast grain boundary model. Contrib Mineral Petrol 112:543-557 Eiler JM, Kitchen N, Rahn TA (2000) Experimental constraints on the stable-isotope systematics of CO2 ice/vapor systems and relevance to the study of Mars. Geochim Cosmochim Acta 64:733-746 Elcombe MM, Hulston JR (1975) Calculation of sulphur isotope fractionation between sphalerite and galena using lattice dynamics. Earth Planet Sci Lett 28:172-180 Emiliani C (1955) Pleistocene paleotemperatures. J Geol 63:538-578 Emrich K, Ehhalt DH, Vogel JC (1970) Carbon isotope fractionation during the precipitation of calcium carbonate. Earth Planet Sci Lett 8:363-371 Epstein S, Buchsbaum R, Lowenstam HA, Urey HC (1953) Revised carbonate-water isotopic temperature scale. Geol Soc Am Bull 64:1315-1326 Epstein S, Graf DL, Degens ET (1963) Oxygen isotope studies on the origin of dolomite. In Isotope and Cosmic Chemistry. H Craig et al. (eds) North-Holland, Amsterdam, p 169-180

Equilibrium Isotopic Fractionation Factors

53

Escande M, Decarreau A, Labeyrie L (1984) Etude experimentale de l’echangeabilite des isotopes de l’oxygene des smectites. C R Acad Sci Paris 299:707-710 Eslinger EV (1971) Mineralogy and oxygen isotope ratios of hydrothermal and low-grade metamorphic argillaceous rocks. PhD Dissertation, Case Western Reserve University, Cleveland, Ohio Eslinger EV, Savin SM (1973) Mineralogy and oxygen isotope geochemistry of the hydrothermally altered rocks of the Ohaki-Broadlands, New Zealand geothermal area. Am J Sci 273:240-267 Eslinger EV, Savin SM, Yeh H (1979) Oxygen isotope geothermometry of diagenetically altered shales. SEPM Spec Publ 26:285-305 Farquhar J (1995) Strategies for high-temperature oxygen isotope thermometry. PhD Dissertation, University of Alberta Farquhar J, Chacko T, Frost BR (1993) Strategies for high-temperature oxygen isotope thermometry: a worked example from the Laramie Anorthosite complex Wyoming USA. Earth Planet Sci Lett 117:407-422 Farver JR (1994) Oxygen self-diffusion in calcite: dependence on temperature and water fugacity. Earth Planet Sci Lett 121:575-587 Fayek M, Kyser TK (2000) Low temperature oxygen isotopic fractionation in the uraninite-UO3-CO2-H2O system. Geochim Cosmochim Acta 64:2185-2197 Feng X, Savin SM (1993) Oxygen isotope studies of zeolites - stilbite, analcite, heulandite, and clinoptilolite; III. Oxygen isotope fractionation between stilbite and water or water vapor. Geochim Cosmochim Acta 57:4239-4247 Fontes JC, Gonfiantini R (1967) Fractionnement istopique de l’hydrogene dans l’eau de cristallisation du gypse. C R Acad Sc Paris 265:4-6 Fortier SM, Lüttge A (1995) An experimental calibration of the temperature-dependence of oxygen isotope fractionation between apatite and calcite at high temperature (350-800°C). Chem Geol 125:281-290 Fortier SM, Lüttge A, Satir M, Metz P (1994) Oxygen isotope fractionation between fluorphlogopite and calcite: An experimental investigation of temperature-dependence and F-/OH- effects. Eur J Mineral 6:53-65 Fortier SM, Cole DR, Wesolowski DJ, Riciputi LR, Paterson BA, Valley JW, Horita J (1995) Determination of equilibrium magnetite-water oxygen isotope fractionation factor at 350°C: a comparison of ion microprobe and laser fluorination techniques. Geochim Cosmochim Acta 59:38713875 Frantz JD, Dubessy J, Mysen B (1993) An optical cell for Raman spectroscopic studies of supercritical fluids and its application to the study of water to 500°C and 2000 bar. Chem Geol 106:9-26 Friedman I, O'Neil JR (1977) Compilation of Stable Isotope Fractionation Factors of Geochemical Interest. U SGeol Surv Prof Paper 440-KK Friedman I, Gleason J, Sheppard RA, Gude AJ, III (1993) Deuterium fractionation as water diffuses into silicic volcanic ash. In Climate Change in Continental Isotopic Records. PK Swart et al. (eds) Am Geophys Union 321-323 Fritz P, Smith DGW (1970) The isotopic composition of secondary dolomites. Geochim Cosmochim Acta 34:1161-1173 Galimov EM (1985) The Biological Fractionation of Isotopes. Academic Press, Orlando, Florida Garlick GD (1966) Oxygen isotope fractionation in igneous rocks. Earth Planet Sci Lett 1:361-368 Gautason B, Chacko T, Muehlenbachs K (1993) Oxygen isotope partitioning among perovskite (CaTiO3), cassiterite (SnO2) and calcite (CaCO3). Geol Assoc Canada/ Mineral Assoc Canada Abstr Programs, Edmonton, Alberta, p A34 Gilg HA, Sheppard SMF (1996) Hydrogen isotope fractionation between kaolinite and water revisited. Geochim Cosmochim Acta 60:529-533 Gillet P, McMillan P, Schott J, Badro J, Grzechnik A (1996) Thermodynamic properties and isotopic fractionation of calcite from vibrational spectroscopy of 18O-substituted calcite. Geochim Cosmochim Acta 60:3471-3485 Giletti BJ (1986) Diffusion effects on oxygen isotope temperatures of slowly cooled igneous and metamorphic rocks. Earth Planet Sci Lett 77:218-228 Giletti BJ, Yund RA (1984) Oxygen diffusion in quartz. J Geophys Res 89B:4039-4046 Girard J-P, Savin SM (1996) Intracrystalline fractionation of oxygen isotopes between hydroxyl and nonhydroxyl sites in kaolinite measured by thermal dehydroxylation and partial fluorination. Geochim Cosmochim Acta 60:469-487 Goldsmith JR (1991) Pressure-enhanced Al/Si diffusion and oxygen isotope exchange. In Ganguly J (ed) Diffusion, Atomic Ordering and Mass Transport. Springer-Verlag, Berlin, p 221-247 Golyshev SI, Padalko NL, Pechenkin, SA (1981) Fractionation of stable oxygen and carbon isotopes in carbonate systems. Geochem Int’l 18:85-99

54

Chacko, Cole & Horita

Gonfiantini R, Fontes JC (1963) Oxygen isotopic fractionation in the water of crystallization of gypsum. Nature 200:644-646 Graham CM, Sheppard SMF, Heaton THE (1980) Experimental hydrogen isotope studies:hydrogen isotope fractionation in the systems epidote-H2O, zoisite- H2O and AlO(OH)-H2O. Geochim Cosmochim Acta 44:353-364 Graham CM, Sheppard SMF (1980) Experimental hydrogen isotope studies, II. Fractionations in the systems epidote-NaCl-H2O, epidote-CaCl2-H2O and epidote-seawater, and the hydrogen isotope composition of natural epidote. Earth Planet Sci Lett 48:237-251 Graham CM, Harmon RS, Sheppard SMF (1984) Experimental hydrogen isotope studies: Hydrogen isotope exchange between amphibole and water. Am Mineral 69:128-138 Graham CM, Viglino JA, Harmon RS (1987) Experimental study of hydrogen-isotope exchange between aluminous chlorite and water and of diffusion in chlorite. Am Mineral 72:566-579 Grinenko VA, Mineyev SD, Devirts AL, Lagutina Ye P (1987) Hydrogen isotope fractionation in the lizardite-water system at 100°C and 1 atm. Geochem Int’l 24:100-104. Grootes PM, Mook WG, Vogel JC (1969) Isotopic fractionation between gaseous and condensed carbon dioxide. Z Physik 221:257-273 Grossman EL, Ku T-L (1986) Oxygen and carbon isotope fractionation in biogenic aragonite: temperature effects. Chem Geol 59:59-74 Gu Z (1980) Determination of the separation facto for isotopic exchange reaction between H2O and CO2 at 25°C. He Huaxue Yu Fangshe Huaxue 2:112-115 (in Chinese) Halas S, Wolacewicz W (1982) The experimental study of oxygen isotope exchange reaction between dissolved bicarbonate and water. J Chem Phys 76:5470-5472 Halas S, Szaran J, Niezgoda H (1997) Experimental determination of carbon isotope equilibrium fractionation between dissolved carbonate and carbon dioxide. Geochim. Cosmochim. Acta 61:26912695 Hamza MS, Epstein S (1980) Oxygen isotopic fractionation between oxygen of different sites in hydroxylbearing silicate minerals. Geochim Cosmochim Acta 44:173-182 Hamza MS, Broecker WS (1974) Surface effect on the isotopic fractionation between CO2 and some carbonate minerals. Geochim Cosmochim Acta 38:669-681 Hariya U, Tsutsumi M (1981) Hydrogen isotopic composition of MnO(OH) minerals from manganese oxide and massive sulfide (Kuroko) deposits in Japan. Contrib Mineral Petrol 77:256-261 Hoering TC (1961) The physical chemistry of isotopic substances: The effect of physical changes in isotope fractionation. Carnegie Inst Washington Yearbook 60:201-204 Hoffbauer R, Hoernes S, Fiorentini E (1994) Oxygen isotope thermometry based on a refined increment method and its application to granulite-grade rocks from Sri Lanka. Precambrian Res 66:199-220 Horibe Y, Craig H (1995) D/H fractionation in the system methane-hydrogen-water. Geochim Cosmochim Acta 59:5209-5217 Horibe Y, Shigehara K, Takakuwa Y (1973) Isotope separation factor of carbon dioxide-water system and isotopic composition of atmospheric oxygen. J Geophys Res 78:2625-2629 Horita J (1989) Stable isotope fractionation factors of water in hydrated saline minerals-brine systems. Earth Planet Sci Lett 95:173-179 Horita J (2001) Carbon isotope exchange in the system CO2-CH4 at elevated temperatures. Geochim Cosmochim Acta 65:1907-1919 Horita J, Wesolowski DJ (1994) Liquid-vapor fractionation of oxygen and hydrogen isotopes of water from the freezing to the critical temperature. Geochim Cosmochim Acta 58:3425-3437 Horita J, Wesolowski DJ, Cole DR (1993a) The activity-composition relationship of oxygen and hydrogen isotopes in aqueous salt solutions: I. Vapor-liquid water equilibration of single salt solutions from 50 to 100°C. Geochim Cosmochim Acta 57:2797-2817 Horita J, Cole DR, Wesolowski DJ (1993b) The activity-composition relationship of oxygen and hydrogen isotopes in aqueous salt solutions: II. Vapor-liquid water equilibration of mixed salt solutions from 50 to 100°C and geochemical implications. Geochim Cosmochim Acta 57:4703-4711 Horita J, Cole DR, Wesolowski DJ (1994) Salt effects on stable isotope partitioning and their geochemical implications for geothermal brines. In Proc 19th Workshop on Geothermal Reservoir Engineering, Stanford University, p 285-290 Horita J, Cole DR, Wesolowski DJ (1995a) The activity-composition relationship of oxygen and hydrogen isotopes in aqueous salt solutions: III. Vapor-liquid water equilibration of NaCl solutions from to 350°C. Geochim Cosmochim Acta 59:1139-1151 Horita J, Wesolowski DJ, Cole DR (1995b) D/H and 18O/16O partitioning between water liquid and vapor in the system H2O-Na-K-Ca-Mg-Cl-SO4 from 0 to 350oC. In Physical Chemistry of Aqueous Systems: Meeting the Needs of Industry. Proc 12th Int’l Conf on the Properties of Water and Steam. HJ White Jr et al. (eds) Begell House, p 505-510

Equilibrium Isotopic Fractionation Factors

55

Horita J, Cole DR, Wesolowski DJ, Fortier SM (1996) Salt effects on isotope partitioning and their geochemical implications: An overview. In Proc Todai Int’l Symp on Cosmochronology and Isotope Geoscience, p 33-36 Horita J, Cole DR, Wesolowski DJ (1997) Salt effects on oxygen and hydrogen isotope partitioning between aqueous salt solutions and coexisting phases at elevated temperatures. In Proc 5th Int’l Symp on Hydrothermal Reactions. Gatlinburg, Tennessee, p 194-197 Horita J, Driesner T, Cole DR (1999) Pressure effect on hydrogen isotope fractionation between brucite and water at elevated temperatures. Science 286:1545 -1547 Horita J, Cole DR, Polyakov VB, Driesner T (in press) Experimental and theoretical study of pressure effects on hydrogen isotope fractionation in the system brucite-water at elevated temperatures. Geochim Cosmochim Acta Hu G, Clayton RN (in press) Oxygen isotope salt effects at high pressure and high temperature, and the calibration of oxygen isotope geothermometers. Geochim Cosmochim Acta James AT, Baker DR (1976) Oxygen isotope exchange between illite and water at 22°C. Geochim. Cosmochim. Acta 40:235-239 Javoy M, Pineau F, Iiyama I (1978) Experimental determination of the isotopic fractionation between gaseous CO2 and carbon dissolved in tholeiitic magma. Contrib Mineral Petrol 67:35-39 Jenkin GRT, Farrow CM, Fallick AE, Higgins D (1994) Oxygen isotope exchange and closure temperatures in cooling rocks. J Metam Geol 12:221-235 Jibao G, Yaqian Q (1997) Hydrogen isotope fractionation and hydrogen diffusion in the tourmaline-water system. Geochim Cosmochim Acta 61:4679-4688 Kakiuchi M (1994) Temperature-dependence of fractionation of hydrogen isotopes in aqueous sodium chloride solutions. J Sol Chem 23:1073-1087 Kakiuchi M (2000) Distribution of isotopic molecules, H2O, HDO, and D2O in vapor and liquid phases in pure water and aqueous solution systems. Geochim Cosmochim Acta 64:1485-1492 Karlsson HR, Clayton RN (1990) Oxygen isotope fractionation between analcime and water: An experimental study. Geochim Cosmochim Acta 54:1359-1368 Kazahaya K (1986) Chemical and isotopic studies on hydrothermal solutions. PhD dissertation, Tokyo Inst Technology Kawabe I (1978) Calculation of oxygen isotope fractionation in quartz-water system with special reference to the low temperature fractionation. Geochim Cosmochim Acta 42:613-621 Kendall C, Chou. I-M, Coplen TB (1983) Salt effect on oxygen isotope equilibria. EOS Trans Am Geophys Union 64:334-335 Kieffer SW (1982) Thermodynamic and Lattice vibrations of Minerals: 4.Application to phase equilibria, isotope fractionation, and high pressure thermodynamics properties. Rev Geophys Space Phys 20:827849 Kim S-T, O’Neil JR (1997) Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochim Cosmochim Acta 61:3461-3475 Kitchen NE, Valley JW (1995) Carbon isotope thermometry in marbles of the Adirondack Mountains, New York. J Metam Geol 13:577-594 Kita I, Taguchi S, Matsubaya O (1985) Oxygen isotope fractionation between amorphous silica and water at 34-93°C. Nature 314:83-34 Koehler G, Kyser TK (1996) The significance of hydrogen and oxygen stable isotopic fractionations between carnallite and brine at low temperature: Experimental and empirical results. Geochim Cosmochim Acta 60:2721-2726 Kohn MJ, Valley JW (1998a) Oxygen isotope geochemistry of amphiboles: isotope effects of cation substitutions in minerals. Geochim Cosmochim Acta 62:1947-1958 Kohn MJ, Valley JW (1998b) Effects of cation substitution in garnet and pyroxene on equilibrium oxygen isotope fractionations. J Metam Geol 16:625-639 Kohn MJ, Valley JW (1998c) Obtaining equilibrium oxygen isotope fractionations from rocks: theory and example. Contrib Mineral Petrol 132:209-224 Kotzer TG, Kyser TK, King RW, Kerrich R (1993) An empirical oxygen- and hydrogen-isotope geothermometer for quartz-tourmaline and tourmaline-water. Geochim Cosmochim Acta 57:34213426 Kulla JB (1979) Oxygen and hydrogen isotope fractionation factors determined in clay-water systems. PhD Dissertation, University of Illinois at Urbana-Champaign Kulla JB, Anderson TF (1978) Experimental oxygen isotope fractionation between kaolinite and water. In Short Papers of the 4th International Congress, Geochronology, Cosmochronology, Isotope Geology. RE Zartman (ed) U S Geol Surv Open file Report 78-70, p 234-235 Kuroda Y, Hariya Y, Suzuoki T, Matsuo S (1982) D/H fractionation between water and the melts of quartz, K-feldspar, albite and anorthite at high temperature and pressure. Geochem J 16:73-78

56

Chacko, Cole & Horita

Kusakabe M, Robinson BW (1977) Oxygen and sulfur isotope equilibria in the BaSO4-HSO4--H2O system from 110 to 350°C and applications. Geochim Cosmochim Acta 41:1033-1040 Kyser TK (1987) Equilibrium fractionation factors for stable isotopes. In Kyser TK (ed) Stable Isotope Geochemistry of Low Temperature Fluids. Mineral Assoc Canada Short Course 13:1-84 Labeyrie L (1974) New approach to surface seawater paleotemperatures using 18O/16O ratios in silica of diatom frustules. Nature 248:40-41 Lambert SJ, Epstein S (1980) Stable isotope investigations of an active geothermal system in Valles Caldera, Jemez Mountains, New Mexico. J Volcan Geotherm Res 8:111-129. Lawrence JR, Taylor HP Jr (1971) Deuterium and oxygen-18 correlation: Clay minerals and hydroxides in Quaternary soils compared to meteoric waters. Geochim Cosmochim Acta 35:993-1003 Lawrence JR, Taylor HP Jr (1972) Hydrogen and oxygen isotope systematics in weathering profiles. Geochim Cosmochim Acta 36:1377-1393 Lecuyer C, Grandjean P, Sheppard SMF (1999) Oxygen isotope exchange between dissolved phosphate and water at temperatures ≤135°C: Inorganic versus biological fractionations. Geochim Cosmochim Acta 63:855-862 Lehmann M, Siegenthaler U (1991) Equilibrium oxygen- and hydrogen-isotope fractionation between ice and water. J Glaciol 37:23-26 Lesniak PM, Sakai H (1989) Carbon isotope fractionation between dissolved carbonate (CO32-) and CO2(g) at 25° and 40°C. Earth Planet Sci Lett 95:297-301 Lichtenstein U, Hoernes S (1992) Oxygen isotope fractionation between grossular-spessartine garnet and water: an experimental investigation. Eur J Mineral 4:239-249 Liu K-K, Epstein S (1984) The hydrogen isotope fractionation between kaolinite and water. Isotope Geosci 2:335-350 Lloyd RM (1968) Oxygen isotope behavior in the sulfate-water system. J Geophys Res 73:6099-6110 Lowenstam HA (1962) Magnetite in denticle capping in recent chitons (polyplacophora). Geol Soc Am Bull 73:435 Majoube M (1971a) Fractionnement en 18O entre la glace et la vapeur d’eau. J Chim Phys 68:625-636 Majoube M (1971b) Fractionnement en oxygene 18 et en deuterium entre l’eau et sa vapeur. J Chim Phys 68:1423-1436 Majzoub M (1966) Une methode d’analyse isotopique de l’oxygene sur des microquantites d’eau determination des coefficients de partage a l’equilibre de l’oxygene 18 entre H2O et CO2, D2O et CO2. J Chim Phys 63:563-568 Malinin SD, Kropotiva OI, Grinenko VA (1967) Experimental determination of equilibrium constants for carbon isotope exchange in the system CO2(gas)-HCO3-(sol) under hydrothermal conditions. Geochem Int’l 4:764-771 Mandernack KW, Bazylinski DA, Shanks WC III, Bullen TD (1999) Oxygen and iron isotope studies of magnetite produced by magnetotactic bacteria. Science 285:1892-1896 Marumo K, Nagasawa K, Kuroda Y (1980) Mineralogy and hydrogen isotope composition of clay minerals in the Ohnuma geothermal area, northeastern Japan. Earth Planet Sci Lett 47:255-262 Matsubaya O, Sakai H (1973) Oxygen and hydrogen isotopic study on the water of crystallization of gypsum from the Kuroko type mineralization. Geochem J 7:153-165 Matsuhisa Y, Matsubaya O, Sakai H (1971) BrF5 technique for the oxygen isotopic analysis of silicates and water. Mass Spectrometer 19:124-133 Matsuhisa Y, Goldsmith JR, Clayton RN (1979) Oxygen isotopic fractionation in the system quartz-albiteanorthite-water. Geochim Cosmochim Acta 43:1131-1140 Matsuo S, Friedman I, Smith GI (1972) Studies of Quaternary saline lakes. I. Hydrogen isotope fractionation in saline minerals. Geochim Cosmochim Acta 36:427-435 Mattey DP (1991) Carbon dioxide solubility and carbon isotope fractionation in basaltic melt. Geochim Cosmochim Acta 55:3467-3473 Mattey DP, Taylor WR, Green DH, Pillinger CT (1990) Carbon isotopic fractionation between CO2 vapour, silicate and carbonate melts: An experimental study to 30 kbar. Contrib Mineral Petrol 104:492-505 Matthews A (1994) Oxygen isotope geothermometers for metamorphic rocks. J Metam Geol 12:211-219 Matthews A, Katz A (1977) Oxygen isotope fractionation during the dolomitization of calcium carbonate. Geochim Cosmochim Acta 41:1431-1438 Matthews A, Schliestedt M (1984) Evolution of blueschist and greenschist rocks of Sifnos, Cylades, Greece. Contrib Mineral Petrol 88:150-163 Matthews A, Beckinsale RD, Durham JJ (1979) Oxygen isotope fractionation between rutile and water and geothermometry of metamorphic eclogites. Mineral Mag 43:405-413 Matthews A, Goldsmith JR, Clayton RN (1983a) Oxygen isotope fractionations involving pyroxenes: The calibration of mineral-pair geothermometers. Geochim Cosmochim Acta 47:631-644

Equilibrium Isotopic Fractionation Factors

57

Matthews A, Goldsmith JR, Clayton RN (1983b) On the mechanisms and kinetics of oxygen isotope exchange in quartz and feldspars at elevated temperatures and pressures. Geol Soc Am Bull 94:396412 Matthews A, Goldsmith JR, Clayton RN (1983c) Oxygen isotope fractionation between zoisite and water. Geochim Cosmochim Acta 47:645-654 Matthews A, Palin, JM, Epstein S, Stolper EM (1994) Experimental study of 18O/16O partitioning between crystalline albite, albitic glass, and CO2 gas. Geochim Cosmochim Acta 58:5255-5266 McCrea JM (1950) On the isotopic chemistry of carbonates and a paleotemperature scale. J Chem Phys 18:849-857 McMillan P (1985) Vibrational spectroscopy in the mineral sciences. In Kieffer SW, Navrotsky A (eds) Microscopic to Macroscopic—Atomic Environments to Mineral Thermodynamics. Rev Mineral 14:963 Melchiorre EB, Criss RE, Rose TP (1999) Oxygen and carbon isotope study of natural and synthetic malachite. Econ Geol 94:245-259 Melchiorre EB, Criss RE, Rose TP (2000) Oxygen and carbon isotope study of natural and synthetic azurite. Econ Geol 95:621-628. Melchiorre EB, Williams PA, Bevins RE (2001) A low temperature oxygen isotope thermometer for cerussite, with applications at Broken Hill, New South Wales, Australia. Geochim Cosmochim Acta 65:2527-2533 Merlivat L, Nief G (1967) Fractionnement isotopique lors de changements d’etat solide-vapeur et liquidevapeur de l’eau a des temperatures inferieures a 0°C. Tellus 19:122-127 Mineev SD, Grinenko VA (1996) The pressure influence on hydrogen isotopes fractionation in the serpentine-water system. V M Goldschmidt Conf Abstr 1:404 Mizutani Y, Rafter TA (1969) Oxygen isotopic composition of sulfate-Part 3. Oxygen isotopic fractionation in the bisulfate ion-water system. New Zealand J Sci 12:54-59 Mook WG, Bommerson JC, Staverman WH (1974) Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth Planet Sci Lett 22:169-176 Morikiyo T (1984) Carbon isotopic study on coexisting calcite and graphite in the Ryoke metamorphic rocks, northern Kiso district. Contrib Mineral Petrol 87:251-259 Morse PM (1929) Diatomic molecules according to the wave mechamics. II. Vibrational levels. Phys Rev 34:57-64 Müller J (1995) Oxygen isotopes in iron (III) oxides: A new preparation line; mineral-water fractionation factors and paleo-environmental considerations. Isotopes Environ Health Stud 31:301-302 Nahr T, Botz R, Bohrmann G, Schmidt M (1998) Oxygen isotopic composition of low-temperature authigenic clinoptilolite. Earth Planet Sci Lett 160:369-381 Northrop DA, Clayton RN (1966) Oxygen-isotope fractionations in systems containing dolomite. J Geol 74:174-196 Noto M, Kusakabe M (1997) An experimental study of oxygen isotope fractionation between wairakite and water. Geochim Cosmochim Acta 61:2083-2093 Ohmoto H (1986) Stable isotope geochemistry of ore deposits. In Valley JW, Taylor HP Jr, O’Neil JR (eds) Stable Isotopes in High Temperature Geological Processes. Rev Mineral 16:491-559 O’Neil, JR (1963) Oxygen isotope fractionation studies in mineral systems. PhD Dissertation, University of Chicago O’Neil JR (1968) Hydrogen and oxygen isotope fractionation between ice and water. J Phys Chem 723:683-3684. O’Neil JR (1986) Theoretical and experimental aspects of isotopic fractionation. In Valley JW, Taylor HP Jr, O’Neil JR (eds) Stable Isotopes in High Temperature Geological Processes. Rev Mineral 16:1-40 O’Neil JR, Clayton RN (1964) Oxygen isotope geothermometry. In Isotope and Cosmic Chemistry (eds., H. Craig et al.) North-Holland, Amsterdam, p157-168 O’Neil JR, Epstein S (1966) A method for oxygen isotope analysis of milligram quantities of water and some of its applications. J Geophys Res 71:4955-4961 O’Neil JR, Taylor HP Jr (1967) The oxygen isotope and cation exchange chemistry of feldspars. J Geophys Res 74:6012-6022 O’Neil JR, Taylor HP Jr (1969) Oxygen isotope equilibrium between muscovite and water. Am Mineral 52:1414-1437 O’Neil JR, Barnes I (1971) C13 and O18 compositions in some fresh-water carbonates associated with ultramafic rocks and serpentines: western United States. Geochim Cosmochim Acta 35:687-697 O'Neil JR, Truesdell AH (1991) Oxygen isotope fractionation studies of solute-water interactions. In Taylor HP Jr, O’Neil JR, Kaplan IR (eds) Stable Isotope Geochemistry: A Tribute to Samuel Epstein, Geochem Soc Spec Pub 3:17-25 O’Neil JR, Clayton RN, Mayeda TK (1969) Oxygen isotope fractionation in divalent metal carbonates. J Chem Phys 51:5547-5558

58

Chacko, Cole & Horita

O’Neil JR, Adami LH, Epstein S (1975) Revised value for the 18O fractionation between CO2 and water at 25°C. U S Geol Surv J Res 3:623-624 Palin JM, Epstein S, Stolper EM (1996) Oxygen isotope partitioning between rhyolitic glass/melt and CO2: An experimental study at 550-950°C and 1 bar. Geochim Cosmochim Acta 60:1963-1973 Patel A, Price GD, Mendelssohn MJ (1991) A computer simulation approach to modelling the structure, thermodynamics and oxygen isotope equilibria of silicates. Phys Chem Minerals 17:690-699 Pineau F, Shilobreeva, S, Kadik A, Javoy M (1998) Water solubility and D/H fractionation in the system basaltic andesite-H2O at 1250°C and between 0.5 and 3 kbars. Chem Geol 147:173-184 Polyakov VB (1997) Equilibrium fractionation of iron isotopes: estimation from Mössbauer spectroscopy data. Geochim Cosmochim Acta 61:4213-4217 Polyakov VB (1998) On anharmonic and pressure corrections to the equilibrium isotopic constants for minerals, Geochim Cosmochim Acta 62:3077-3088 Polyakov VB, Kharlashina NN (1994) Effect of pressure on equilibrium isotopic fractionation. Geochim Cosmochim Acta 58:4739-4750 Polyakov VB, Kharlashina NN (1995) The use of heat capacity data to calculate carbon dioxide fractionation between graphite, diamond, and carbon dioxide: A new approach. Geochim Cosmochim Acta 59:2561-2572 Polyakov VB, Mineev SD (2000) The use of Mössbauer spectroscopy in stable isotope geochemistry. Geochim Cosmochim Acta 64:849-865 Poulson SR, Schoonen MAA (1994) Variations of the oxygen isotope fractionation between NaCO3- and water due to the presence of NaCl at 100-300°C. Chem Geol (Isotope Geosci Sec) 116:305-315 Pradahananga TM, Matsuo S (1985a) Intracrystalline site preference of hydrogen isotopes in borax. J Phys Chem 89:72-76 Pradahananga TM, Matsuo S (1985b) D/H fractionation in sulfate hydrate-water systems. J Phys Chem 89:1869-1872 Richet P, Bottinga Y, Javoy M (1977) A review of hydrogen, carbon, nitrogen, oxygen, sulphur, and chlorine stable isotope fractionation among gaseous molecules. Ann Rev Earth Planet Sci 5:65-110 Richet P, Roux J, Pineau F (1986) Hydrogen isotope fractionation in the system H2O-liquid NaAlSi3O8: new data and comments on D/H fractionation in hydrothermal experiments. Earth Planet Sci Lett 78:115-120 Richter R, Hoernes S (1988) The application of the increment method in comparison with experimentally derived and calculated O-isotope fractionations. Chem Erde 48:1-18 Rolston JH, Hartog J den, Butler JP (1976) The deuterium isotope separation factor between hydrogen and liquid water. J Phys Chem 80:1064-1067 Romanek CS, Grossman EL, Morse JW (1992) Carbon isotopic fractionation in synthetic aragonite and calcite: Effects of temperature and precipitation rate. Geochim Cosmochim Acta 56:419-430 Rosenbaum JM (1993) Room temperature oxygen isotope exchange between liquid CO2 and H2O. Geochim Cosmochim Acta 57:3195-3198 Rosenbaum JM (1994) Stable isotope exchange between carbon dioxide and calcite at 900°C. Geochim Cosmochim Acta 58:3747-3753 Rosenbaum JM (1997) Gaseous, liquid and supercritical H2O and CO2: oxygen isotope fractionation behavior. Geochim Cosmochim Acta 61:4993-5003 Rosenbaum JM, Mattey DP (1995) Equilibrium garnet-calcite oxygen isotope fractionation. Geochim Cosmochim Acta 59:2839-2842 Rosenbaum JM, Kyser TK, Walker D (1994) High temperature oxygen isotope fractionation in the enstatite-olivine-BaCO3 system. Geochim Cosmochim Acta 58:2653-2660 Rowe MW, Clayton RN, Mayeda TK (1994) Oxygen isotopes in separated components of CI and CM meteorites. Geochim Cosmochim Acta 58:5341-5347 Rubinson M, Clayton RN (1969) Carbon-13 fractionation between aragonite and water. Geochim Cosmochim Acta 33:997-1002 Rye RO, Stoffregen RE (1995) Jarosite-water oxygen and hydrogen isotope fractionations: Preliminary experimental data. Econ Geol 90:2336-2342 Rye RO, Bethke PM, Wasserman MD (1992) The stable isotope geochemistry of acid-sulfate alteration. Econ Geol 87:240-262 Saccocia PJ, Seewald JS, Shank WC III (1998) Hydrogen and oxygen isotope fractionation between brucite and aqueous NaCl solutions from 250-450°C. Geochim Cosmochim Acta 62:485-492 Sakai H, Tsutsumi M (1978) D/H fractionation factors between serpentine and water at 10° to 500°C and 2000 bar water pressure, and the D/H ratios of natural serpentine. Earth Planet Sci Lett 40:231-242 Satake H, Matsuo S (1984) Hydrogen isotopic fractionation factor between brucite and water in the temperature range from 100 to 510°C. Contrib Mineral Petrol 86:19-24 Sato RK, McMillan PF (1987) Infrared spectra of the isotopic species of alpha quartz. J Phys Chem 91:3494-3498

Equilibrium Isotopic Fractionation Factors

59

Savin SM, Lee M (1988) Isotopic studies of phyllosilicates. In Bailey SW (ed) Hydrous Phyllosilicates (exclusive of micas). Rev Mineral 19:189-219 Scheele N, Hoefs J (1992) Carbon isotope fractionation between calcite, graphite and CO2: An experimental study. Contrib Mineral Petrol 112:35-45 Schauble E, Rossman GR, Taylor, HP (2001) Theoretical estimates of equilibrium Fe-isotope fractionations from vibrational spectroscopy. Geochim Cosmochim Acta 65:2487-2497 Schütze H (1980) Der Isotopenindex—eine Inkrementenmethode zur näherungsweisen Berechnung von Isotopenaustauschgleichgewichten zwischen kristallinen Substanzen. Chem Erde 39:321-334 Schwarcz HP (1966) Oxygen isotope fractionation between host and exsolved phases in perthite. Geol Soc Am Bull 77:879-882 Sharp ZD (1995) Oxygen isotope geochemistry of the Al2SiO5 polymorphs. Am J Sci 295:1-19 Sharp ZD, Kirschner DL (1994) Quartz-calcite oxygen isotope thermometry: a calibration based on natural isotopic variations. Geochim Cosmochim Acta 58:4491-4501 Sheppard SMF, Schwarcz HP (1970) Fractionation of carbon and oxygen isotopes and magnesium between coexisting metamorphic calcite and dolomite. Contrib Mineral Petrol 26:161-198 Sheppard SMF, Gilg HA (1996) Stable isotope geochemistry of clay minerals. Clay Minerals 31:1-24 Sheppard SMF, Nielsen RL, Taylor HP Jr (1969) Oxygen and hydrogen isotope ratios of clay minerals from porphyry copper deposits. Econ Geol 64:755-777 Shilobreyeva SN, Devirts AL, Kadik AA, Lagutina YP (1992) Distribution of hydrogen isotopes in basalt liquid-water equilibrium at 3 kbar and 1250°C. Geochem Int’l 29:130-134 Shmulovich K, Landwehr D, Simon K, Heinrich W (1999) Stable isotope fractionation between liquid and vapor in water-salt systems up to 600°C. Chem Geol 157:343-354 Smyth JR (1989) Electrostatic characterization of oxygen sites in minerals. Geochim Cosmochim Acta 53:1101-1110 Smyth JR, Clayton RN (1988) Correlation of oxygen isotope fractionations and electrostatic site potentials in silicates. EOS Trans Am Geophys Union 69:1514 Sofer Z (1978) Isotopic composition hydration water in gypsum. Geochim Cosmochim Acta 42:1141-1149 Sofer Z, Gat JR (1972) Activities and concentrations of oxygen-18 in concentrated aqueous salt solutions: Analytical and geophysical implications. Earth Planet Sci Lett 15:232-238 Sofer Z, Gat JR (1975) The isotope composition of evaporating brines: Effects of the isotopic activity ratio in saline solutions. Earth Planet Sci Lett 26:179-186 Sommer MA, Rye D (1978) Oxygen and carbon isotope internal thermometry using benthic calcite and aragonite foraminifera pairs. Short Papers, 4th Int’l Conf. Geochron Cosmochron Isotope Geol, U S Geol Surv Open File Rep. 78-701, p 408-410 Spindel W, Stern MJ, Monse EU (1970) Further study on temperature-dependences of isotope effects. J Chem Phys 52:2022-2035 Staschewski D (1964) Experimentelle bestimmung der O18/O16-trennfaktoren in den systemen CO2/H2O und CO2/D2O. Berichte Bunsengesell 68:454-459 Stern MJ, Spindel W, Monse EU (1968) Temperature-dependences of isotope effects. J Chem Phys 48:29082919 Stewart MK (1974) Hydrogen and oxygen isotope fractionation during crystallization of mirabilite and ice. Geochim Cosmochim Acta 38:167-172 Stewart MK, Friedman I (1975) Deuterium fractionation between aqueous salt solutions and water vapor. J Geophys Res 80:3812-3818 Stoffregen RE, Rye RO, Wasserman MD (1994) Experimental studies of alunite: I. 18O-16O and D-H fractionation factors between alunite and water at 250-450°C. Geochim Cosmochim Acta 58:903-916 Stolper E, Epstein S (1991) An experimental study of oxygen isotope partitioning between silica glass and CO2 vapor. In Taylor HP Jr, O’Neil JR, Kaplan IR (eds) Stable Isotope Geochemistry: A Tribute to Samuel Epstein. Geochem Soc Spec Pub 3:35-51 Suess, VH (1949) Das gleichgewicht H2 + HDO = HD + H2O und die weiteren Austauschgleichgewichte im System H2, D2, und H2O. Z Naturforschung 4a:328-332 Sushchevskaya TM, Ustinov VI, Nekrasov IY, Gavrilov YY, Grinenko VA (1985) The oxygen-isoope fractionation factor in cassiterite synthesis. Geochem Int’l 23:57-60 Suzuoki T, Kimura T (1973) D/H and 18O/16O fractionation in ice-water system. Mass Spectrom 21:229-233 Suzuoki T, Epstein S (1976) Hydrogen isotope fractionation between OH-bearing minerals and water. Geochim Cosmochim Acta 40:1229-1240 Szaran J (1997) Achievement of carbon isotope equilibrium in the system HCO3- (solution)-CO2 (gas). Chem Geol 142:79-86 Tarutani T, Clayton RN, Mayeda TK (1969) The effect of polymorphism and magnesium substitution on oxygen isotope fractionation between calcium carbonate and water. Geochim Cosmochim Acta 33:987-996

60

Chacko, Cole & Horita

Taube H (1954) Use of oxygen isotope effects in the study of hydration of ions. J Phys Chem 58:523-528 Taylor BE (1976) Origin and significance of C-O-H fluids in the formation of Ca-Fe-Si skarn, Osgood Mountains, Humboldt County, Nevada. PhD Dissertation, Stanford University Taylor BE, Westrich HR (1985) Hydrogen isotope exchange and water solubility in experiments using natural rhyolite obsidian. EOS Trans Am Geophys Union 66:387. Taylor HP Jr (1974) The application of oxygen and hydrogen isotope studies to problems of hydrothermal alteration and ore deposition. Econ Geol 69:843-883 Taylor HP Jr, Epstein S (1962) Relationship between O18/O16 ratios in coexisting minerals of igneous and metamorphic rocks. Part I. Principles and experimental results. Geol. Soc Am Bull 73:461-480 Tennie A, Hoffbauer R, Hoernes S (1998) The oxygen isotope fractionation behaviour of kyanite in experiment and nature. Contrib Mineral Petrol 133:346-355 Truesdell AH (1974) Oxygen isotope activities and concentrations in aqueous salt solutions at elevated temperatures: Consequences for isotope geochemistry. Earth Planet Sci Lett 23:387-396 Turner JV (1982) Kinetic fractionation of carbon-13 during calcium carbonate precipitation. Geochim Cosmochim Acta 46:1183-1191 Urey HC (1947) The thermodynamic properties of isotopic substances. J Chem Soc (London), p 562-581 Usdowski E, Hoefs J (1993) Oxygen isotope exchange between carbonic acid, bicarbonate, carbonate, and water: A re-examination of the data of McCrea (1950) and an expression for the overall partitioning of oxygen isotopes between the carbonate species and water. Geochim Cosmochim Acta 57:3815-3818 Usdowski E, Michaelis J, Bottcher ME, Hoefs J (1991) Factors for the oxygen isotope equilibrium fractionation between aqueous and gaseous CO2, carbonic acid, bicarbonate, carbonate, and water (19°C). Z Phys Chemie 170:237-249 Ustinov VI, Grinenko VA (1990) Determining isotope-equilibrium constants for mineral assemblages. Geochem Int’l 27 (10):1-9 Valley JW, O’Neil JR (1981) 13C/12C exchange between calcite and graphite: a possible thermometer in Grenville marbles. Geochim Cosmochim Acta 45:411-419 Valley JW, Taylor HP Jr, O’Neil JR (eds) (1986) Stable Isotopes in High Temperature Geological Processes. Rev Mineral, Vol 16 Valley JW, Chiarenzelli JR, McLelland JM (1994) Oxygen isotope geochemistry of zircon. Earth Planet Sci Lett 126:187-206 Van Hook WA (1975) Condensed phase isotope effects, especially vapor pressure isotope effects: aqueous solutions. In Rock PA (ed) Isotopes and Chemical Principles. Am Chem Soc Symp 11:101-130. Vennemann TW, O’Neil JR (1996) Hydrogen isotope exchange between hydrous minerals and molecular hydrogen: I. A new approach for the determination of hydrogen isotope fractionation at moderate temperature. Geochim Cosmochim Acta 60:2437-2451 Vitali F, Longstaffe FJ, Bird MI, Caldwell WGE (2000) Oxygen-isotope fractionation between aluminumhydroxide phases and water at 4 hr. Direct exchange at P=100-200 bar. Large number of measurements at room temperature. Brenninkmeijer et al. combined their own and earlier experimental results. Fractionations adjusted such that D(CO2-H2O) at 25°C = 1.0412.

Freezing from 2.5m NaCl soln. Slow growth of ice in a stirred water. Slow distillation of liquid water.

Slow growth of ice in a stirred water. Freezing from 4‰ chlorinity and seawater.

Comments

Appendix 2. Oxygen isotope fractionation factors: calibrations based on experiments or natural samples.

64 Chacko, Cole & Horita

O’Neil et al. (1969)

Clayton et al. (1975)

calcite-H2O

calcite-H2O

Ex

McCrea (1950)

31

36 O’Neil et al. (1969)

Ex

Ex

Ex

O’Neil et al. (1969)

cerrusite (PbCO3) H 2O rhodochrosite (MnCO3) - H2O otavite (CdCO3) H 2O

34 O’Neil et al. (1969)

Ex

O’Neil et al. (1969)

witherite (BaCO3) H 2O

33

35

Ex

O’Neil et al. (1969)

strontianite (SrCO3)H 2O

Ex

Ex

32

29

Mx

Ex

Epstein et al. (1953)

Ex

Gonfiantini and Fontes (1963), and Matsubaya and Sakai (1973), and Sofer (1978) Stewart (1974)

calcite-H2O

Ex

Horita (1989)

30

Ex

Ex

Ex

Horita (1989)

Koehler and Kyser (1996)

Horita (1989)

mirabilite (Na2SO4. 10H2O)- H2O Carbonates: aragonite/calciteH 2O

Hydrated Salts: carnallite(KMgCl3. 6H2O) – H2O carnallite(KMgCl3. 6H2O) – H2O bischofite(MgCl2. 6H2O) – H2O tachyhydrite(CaMg2 Cl6. 12H2O) – H2O gypsum (CaSO4. 2H2O)- H2O

28

27

26

25

24

23

2.95, 5.97

6.76

1.22 - 1.33 (500°) 0.01 - 0.10 (700°C) 2.69 (106/T2) - 3.24 corrected in Friedman and O’Neil (1977) 2.57 (106/T2) - 4.23 corrected in Friedman and O’Neil (1977) 4.51, 6.92

2.78 (106/T2) – 2.89 corrected in Friedman and O’Neil (1977)

-1.2 to 79.8

G18O=15.7 (103/T) - 54.2 (Florida seawater) G18O=16.4 (103/T) - 57.6 (Cape Cod seawater) 2.73 (106/T2) - 2.71

320, 250

240

240, 201

0-500

0-500

500, 700

0-500

7-30

0-25

17-57

25

10-40

22-45

10-40

+1.4 to +2.0

+2.9 to +4.1

+9.5

+7.6 to +8.2

-0.1 to +1.3

+7.5 to +8.8

As above.

As above.

No equation given.

As above.

As above.

Combination of data obtained from biogenic precipitation of calcite in tank experiments and analysis of natural samples. Regression line fit through data given in Epstein et al. (1953) after recalculation following the method outlined in Tarutani et al. (1969). 200-500°C experiments involved exchange between carbonate minerals and ammonium chloride solutions. P=1 kbar. Experiments at 0 and 25 involved controlled precipitation of carbonate minerals from bicarbonate solutions. P= 1 atm. 100% exchange in all experiments. Pressure effects investigated to 20 kbars in pure water.

Slow precipitation by CO2 degassing.

Synthesis from aqueous solutions. Determined on the composition scale of brine.

By precipitation and aging 30-67 days. Determined on the activity scale of brines. By precipitation and aging 30-67 days. Determined on the activity scale of brines. Slow precipitation.

By precipitation and aging 30-67 days. Determined on the activity scale of brines. By synthesis and exchange.

Equilibrium Isotope Fractionation Factors 65

Clayton et al. (1968) Northrop and Clayton (1966) Matthews and Katz (1977) Fritz and Smith (1970) O’Neil and Epstein (1966) O’Neil and Epstein (1966) Chacko et al. (1991)

Scheele and Hoefs (1992)

dolomite-H2O dolomite-H2O

protodolomite-H2O

CO2-calcite

CO2-dolomite CO2-calcite

CO2-calcite

48 49

50

51

52 53

54

Ex

O’Neil et al. (1971)

Ex

Ex Ex

Ex

Ex

Ex Ex

N

Ex

Ex

Melchiorre et al. (2001) Bottcher (2000)

Ex

Ex

Melchiorre et al. (1999) Melchiorre et al. (2000)

Ex

Zhang et al. (2001)

Ex

47

46

45

44

43

42

41

Carothers et al. (1988)

Ex

Kim and O’Neil (1997)

otavite (CdCO3) H 2O siderite (FeCO3)H 2O siderite (FeCO3)H 2O malachite (CuCO3Cu(OH)2)H 2O azurite (CuCO3)2Cu(OH)2)H 2O cerussite(PbCO3)H 2O norsethite (BaMg(CO3)2) H 2O hydromagnesite(Mg 4(OH)4(CO3)3)- H2O dolomite-H2O

39

40

Ex Ex

Kim and O’Neil (1997) Kim and O’Neil (1997)

calcite-H2O witherite (BaCO3) H 2O

37 38

33-197 45-75 0-50 10-45 20-65 20-90

3.13(106/T2) - 3.50 2.56 (106/T2) + 1.69 2.66 (106/T2) + 2.66 2.67 (106/T2) + 4.75 2.63 (106/T2) - 3.58 2.83 (106/T2) - 2.85

1.31 (106/T2) + 3.62 -0.038435 + 5.0077x - 1.0703x2 + 0.15452x3 – 0.014366x4 + 0.00073624x5 – 0.000015567x6 , where x=106/T2 5.92-2.31

500-1200

350-610 400-800

350-610

25-78.6

300-510 252-295

3.20 (106/T2) - 2.00 3.06 (106/T2) - 3.24 +23.4 to +31.6 3.2 (106/T2) – 2.0 1.93 (106/T2) + 3.92

20±5

+34.5

0, 25

0-500

2.76 (106/T2) - 3.96

+31.19, +37.08

10-40 0-500

18.03 (103/T) - 32.42 2.63 (106/T2) - 4.04

Aragonite starting material inverted to calcite during the experiment. No equation given but fractionations generally larger than those given by Chacko et al. (1991) by ~0.5‰.

Based on sedimentary dolomite from Deep Springs Lake, California. Direct exchange. 3-50% exchange. P= 1 kbar. Hydrothermal dolomitization of calcite or aragonite in the presence of Ca-Mg-Sr chloride solutions. P= 1 atm. Precipitated from a Ca-Mg-CO3 soln. The equation is based on extrapolation to the 25°C datum of Clayton et al. (1968). Direct exchange experiments using very large calcite to CO2 ratio. Measured values may represent surface fractionations rather than true equilibrium fractionations. P=0.3 bars. As above. Direct exchange. 39-94% exchange. P=10 kbar. Equation represents theoretical calculations that closely fit the experimental data. The equation reproduces the calculated fractionations from 273-4000K.

Slow precipitation from Mg(HCO3)2 soln.

Formed from BaCO3 and MgCO33H2O in NaHCO3 soln.

Replacement of calcite with Pb2+-bearing soln.

Replacement of calcite with Cu2+-bearing soln.

Based on microbial siderite precipitated by thermophilic Fe(III)-reducing bacteria in culture. Replacement of calcite with Cu2+-bearing soln.

Slow addition of FeCl2 soln to NaHCO3 soln.

Low-T controlled precipitation experiments. Low-T controlled precipitation experiments combined with the high-T experiments of O’Neil et al. (1969). Revised acid fractionation factors used in recalculating the isotopic compositions of BaCO3 determined in O’Neil et al. (1969). As above.

66 Chacko, Cole & Horita

Ex N

Clayton et al. (1989) Sharp and Kirschner (1994) Clayton et al. (1972) Bottinga and Javoy (1973) Matsuhisa et al. (1979) Matsuhisa et al. (1979) Matthews et al. (1983b) Zhang et al. (1989) Kita et al. (1985) Labeyrie (1974)

dolomite-calcite dolomite-calcite

Silica Group: quartz-calcite quartz-calcite quartz-H2O quartz- H2O

quartz-H2O quartz-H2O quartz-H2O

quartz-H2O

amorphous silicaH 2O biogenic silica-H2O

biogenic silica-water

quartz-microcline

quartzcassiterite(SnO2) Feldspars: albite-calcite albite-H2O

60 61

62 63 64 65

66 67 68

69

70

72

73

74

75 76

71

Ex

O’Neil and Epstein (1966) Sheppard and Schwarz (1970)

dolomite-calcite dolomite-calcite

Ex Ex

Ex

Zhang et al. (1994) Clayton et al. (1989) O’Neil and Taylor (1967)

Ex

Ex

Ex Ex Ex

Ex N Ex Mx

Ex N

Ex Ex

Blattner and Bird (1974)

Brandriss et al. (1998)

Epstein et al. (1963) Northrop and Clayton (1966)

N

58 59

Grossman and Ku (1986)

aragonite-calcite

Ex Ex

57

Rosenbaum (1994) Tarutani et al. (1969)

CO2-calcite aragonite-calcite

55 56

400, 500 600-800 350-800

-0.56 (106/T2) 2.91 (106/T2) - 3.41

600 +8.5, +6.9

1.8

3.6-20.0

3.52 (106/T2) - 4.35

15.56 (103/T)-20.92

34-93

3.306 (106/T2) – 2.71

4-27

500-800 250-500 600, 400, 250 180-550

2.05 (106/T2) - 1.14 3.34 (106/T2) - 3.31 1.44, 3.99, 9.12

41.2 - 0.25T(°C)

600-1000 100-700 500-750 500-800

350-610 100-650

0.56 (106/T2) + 0.45 0.45 (106/T2) - 0.40

0.38 (106/T2) 0.87 (106/T2) 2.51 (106/T2) - 1.96 4.10 (106/T2) - 3.7

550 300-510

0-25

900 25

+0.9 0. 50 (106/T2)

0.76 - 0.017T(°C)

3.30 0.6

Direct exchange. 88-100% exchange. P=11-16 kbar. Alkali exchange with aqueous chloride solutions. Found no difference in the fractionation behavior of Na- and Kfeldspar. 100% exchange. P=1 kbar.

Direct exchange. 72-99% exchange. P=15 kbar. Primarily based on analyses of low-grade marbles and veins. Direct exchange. 100% exchange. P=1 kbar. Based on selected experimental data from Clayton et al. (1972), theoretical considerations and data from natural samples. Direct exchange. 87-100% exchange. P=15 kbar. Direct exchange. 36-87% exchange. P=15 kbar. Three isotope technique. No equation given but data fits equation of Matsuhisa et al (1979). Conversion of silica gel to quartz in up to 40wt-% NaCl, NaF, and KCl. Little salt effect observed. Based on the analysis of amorphous silica precipitating in geothermal power plant waters. Based on the analysis of sponge spicules and diatoms formed under known conditions. Based on cultured fresh water diatoms. 3-4‰ smaller than Labeyrie (1974). Quartz and alkali feldspar directly exchanged in the presence of a common water or KCl solution. Reaction between silica gel and amorphous SnO2 in the presence of water.

Direct exchange. 97% exchange. P=12.5 kbar Combination of data from experiments in which calcite or aragonite were slowly precipitated from aqueous bicarbonate solutions. Based on the analysis of aragonitic foraminifer Hogelundia and coexisting calcitic foraminifer Uvigerina. Large scatter about the regression line indicates that the apparent temperature dependence is not statistically significant. Dolomitization of natural calcite with CaCl2 soln Combination of dolomite-H2O experiments of Northrop and Clayton (1966) and calcite-H2O experiment of O’Neil (1963). Combination of CO2-calcite and CO2-dolomite experiments. Based on analysis of co-existing calcite-dolomite pairs in regionally metamorphosed marbles and calcareous schists. Temperatures derived from calcite-dolomite solvus thermometry.

Equilibrium Isotope Fractionation Factors 67

Bottinga and Javoy (1973) Matsuhisa et al. (1979) Matsuhisa et al. (1979) Matthews et al. (1983b) Chiba et al. (1989) Zheng et al. (1994a) Rosenbaum et al. (1994)

anorthite-H2O

anorthite-H2O anorthite-H2O anorthite-H2O

Olivine: forsterite-calcite forsterite-calcite

forsterite-BaCO3

83

84 85 86

87 88

89

grossular- H2O

grossular-andradite

garnet-quartz

94

95

96

91 92 93

Garnet: grossular/andraditecalcite grossular-H2O andradite-H2O grossular/spessartine -H2O

Ex

Clayton et al. (1989) O’Neil and Taylor (1967)

anorthite-calcite anorthite-H2O

81 82

90

Ex Ex

Matsuhisa et al. (1979) Matthews et al. (1983b)

albite-H2O albite-H2O

79 80

Bottinga and Javoy (1975)

Kohn and Valley (1998b) N

N

Ex

Ex Ex Ex

Taylor (1976) Taylor (1976) Lichtenstein and Hoernes (1992)

Matthews (1994)

Ex

Rosenbaum and Mattey (1995)

Ex Ex Ex

Mx

Ex Ex

Ex Ex

Ex

Matsuhisa et al. (1979)

albite-H2O

78

Mx

Bottinga and Javoy (1973)

albite-H2O

77

400-500 600, 500 600-800 500-800 500-800 500-750 400-500 600

2.39 x106/T2-2.51 0.95, 1.38 -1.59 (106/T2) 2.15 (106/T2) - 3.82 2.09 (106/T2) 2- 3.70 1.04 (106/T2) - 2.01 1.49 (106/T2) - 2.81 -1.31

727 >500

-2.88 (106/T2)

700, 800

600 600 750

800-1200

1009-1409 in 100°C increments

0.6 to 0.8

-1.77, -1.51

-1.6 -3.28 -2.1

-2.77 (106/T2)

-1.95, -1.7, -2.12, -1.32, -1.16

700-1300 600-900

500-700

1.59 (106/T2) - 1.16

-3.29 (106/T2) -3.17 (106/T2) - 0.44

500-800

3.13 (106/T2) - 3.7

Direct exchange between grossular0.7 andradite0.19 pyrope0.03 garnet and calcite. 60-97% exchange. P=23 kbar. Hydrothermal synthesis of grossular from gel. P= 2 kbar. Hydrothermal synthesis of andradite from gel. P= 2 kbar. Direct exchange between grossular and spessartine rich garnet and water. Both sets of experiments yielded nearly identical fractionation factors. 56-59% exchange. P=16 kbar. Synthesis experiments at 750°C yield a spessartineH2O fractionation of –2.5. Direct exchange of grossular with 0.86M NaF solution. 9092% exchange. P=1.6-2.0 kbar. No equation given. Based on analysis of coexisting garnet-wollastonite pairs in granulite-facies calc-silicates. Based on quartz-garnet data on natural samples where temperatures for those samples were determined by the Bottinga and Javoy (1973) calibrations of the feldsparquartz, feldspar-muscovite or feldspar-magnetite isotope thermometers.

Direct exchange. 61-100% exchange. P= 15-16 kbar. Exchange of forsterite and calcite in the presence of a CO2H2O fluid. 60-90% exchange. P=3-12 kbar. No equation given but all points are similar to forsteritecalcite fractionations of Chiba et al. (1989) except the 1209°C datum.

Direct exchange. 95-100% exchange. P=9-12 kbar. Exchange of Ba feldspar or anorthite with CaCl2 solutions. 96-100% exchange. P=1 kbar. Based on experimental data from O’Neil and Taylor (1967), theoretical considerations and data from natural samples. Direct exchange. 100% exchange. P=4-10 kbar. Direct exchange. 95-100% exchange. P=2-4 kbar. Three isotope technique. Less than 50% exchange. P=7 kbar.

Based on experimental data from O’Neil and Taylor (1967), theoretical considerations and data from natural samples. Direct exchange. 79-100% exchange. P=7-15 kbar. P=7-12 kbar. Direct exchange. 50-79% exchange. P=12 kbar. Three isotope technique. No equation given but data fits equation of Matsuhisa et al (1979). P=15 kbar.

68 Chacko, Cole & Horita

Tennie et al. (1998)

Sharp (1995) Matthews et al. (1983c)

Kotzer et al. (1993)

Chiba et al. (1989) Rosenbaum et al. (1994) Matthews et al (1983a) Matthews et al (1983a) Matthews et al (1983a) Matthews et al (1983a) Zheng et al. (1994b)

Aluminosilicate kyanite-calcite

kyanite-quartz

sillimanite-quartz

Sorosilicates zoisite-H2O

epidote-quartz

gehlenite-calcite Cyclosilicates: tourmaline-quartz

Pyroxenes: diopside-calcite enstatite- BaCO3

diopside-H2O

wollastonite- H2O

hedenbergite- H2O

jadeite- H2O

Amphiboles: tremolite-calcite

98

99

100

101

102

103

104

105 106

107

108

109

110

111

Chacko et al. (1989)

Matthews and Schliestedt (1984)

Sharp (1995)

Valley et al. (1994)

garnet-zircon

97

Ex

Ex

Ex

Ex

Ex

Ex Ex

N

Ex

Ex

Ex

N

N

Ex

N

-3.80 (106/T2) + 1.67

0.37, 0.21

-1.00

-1.41, -1.37, -1.19, -1.03

-1.27, -1.08, -0.98

520-680

500, 600

600-1200 1009-1409 in 100°C increments 600, 700, 800 500, 600, 700, 800 700

200-600

-1.0 (106/T2) – 0.39

-2.37 (106/T2) -0.97, -1.20, -1.5, -1.2, 1.11

700-1000

400, 450, 500, 600, 700

-3.12 (106/T2)

- (1.56 + 1.92EPs)(106/T2)

0.49, 0.20, -0.33, -0.5, -0.62

535-1300

-2.36 (106/T2)

625-725 535-1300

2

800-1000

-2.17 (10 /T )

6

-2.62 (106/T2)

0

Exchange of tremolite and calcite in the presence of a CO2H2O fluid. 48-81% exchange. P=3-10 kbar. Extrapolation to equilibrium fractionations at 520, 560, 580°C are suspect because of unequal exchange rates in companion experiments at each of those temperatures.

Direct exchange. 61-100% exchange. P=15-16 kbar. Direct exchange. 91-100% exchange. P= 30 kbar. No equation given. Experimental data are non-linear with respect to 1/T2 even at these high temperatures. Direct exchange using three-isotope method. 55- 79% exchange. P=13-18 kbar. No equation given. Direct exchange using three-isotope method. 46-100% exchange. P=9-20 kbar. No equation given. Direct exchange using three-isotope method. 52-73% exchange. P=13 kbar. No equation given. Direct exchange using three-isotope method. 20-39% exchange. P=16-18 kbar. No equation given.

Based on the analysis of co-existing quartz, muscovite and tourmaline from several ore deposits. Temperatures based on quartz-muscovite fractionations compiled by Eslinger et al. (1979).

Direct exchange at 600 and 700°C using the three isotope method. 54-65% exchange. Synthesis from glass in 400600°C experiments. Results from direct exchange and synthesis experiments agree at 600. No equation given. Based on the experimental zosite-water and quartz-water calibrations of Matthews et al. (1983b) and Matsuhisa et al. (1979). Effect of Fe3+ substitution is estimated using the grossular/andradite data of Taylor (1976). EPs is the mole fractionation of pistacite (Ca2Fe3Si3O12OH) component in the epidote. Direct exchange. 90-100% exchange. P=15 kbar.

Isotopic exchange induced by polymorphic inversion of andalusite to kyanite in the presence of calcite. P=13 kbar. Based on the analysis of coexisting quartz, kyanite and garnet and an assumed '(qtz-grt) = 3.1x106/T2. Based on the analysis of coexisting quartz, sillimanite and garnet and an assumed '(qtz-grt) = 3.1x106/T2.

Based on analysis of garnet-zircon pairs from the Adirondack mountains.

Equilibrium Isotope Fractionation Factors 69

O’Neil and Taylor (1969) Bottinga and Javoy (1973)

Bottinga and Javoy (1975)

Wenner and Taylor (1971) Cole and Ripley (1999)

Eslinger (1971) Kulla and Anderson (1978)

phlogopite-calcite

fluorophlogopitecalcite fluorophlogopitecalcite

muscovite-H2O

muscovite-H2O

muscovite-quartz

biotite-quartz

Other Hydrous Phyllosilicates: chlorite- H2O

chlorite- H2O

kaolinite- H2O

kaolinite- H2O

114

115

117

118

119

120

121

122

123

124

116

Chacko et al. (1996)

Micas: muscovite-calcite

113

Matthews and Schliestedt (1984)

Fortier et al. (1994)

Chacko et al. (1996)

Chacko et al. (1996)

Bottinga and Javoy (1975)

amphibole-quartz

112

Ex

Mx

Ex

N

N

Ex

Mx

Ex

Ex

Ex

Ex

Ex

N

500-800 400-800

400-650 500-800 500-650

-1.26 (106/T2) -1.84 (106/T2) + 0.43

2.38 (106/T2) - 3.89 1.90 (106/T2) - 3.10 -1.55 (106/T2)

150-400

0-350 170-320

2.50 (106/T2) - 2.87 2.42 (106/T2) - 4.45

3

2.693 (10 /T ) - 6.342 (106/T2) + 2.969 (103/T)

9

1.56 (106/T2) - 4.70

150-400

650-800

-1.78 (106/T2)

-3.69 (106/T2) + 0.60

550-650

>500

-0.99 (106/T2)

-3.15 (106/T2) + 0.30

Based on the analysis of natural chlorites and coexisting minerals in metasediments. Direct exchange at 350°C. 12% exchange. P=0.25 kbar. This high temperature datum is combined with results from granite-fluid experiments at 170-300°C in which chlorite is formed by alteration of biotite. Combination of single sample of hydrothermal kaolinite from Broadlands, New Zealand and model calculations. Hydrothermal synthesis of kaolinite from gels.

Direct exchange. 80-100% exchange. P=15 kbar. Linear equation fit through the origin. Equation should not be extrapolated below 500°C. Direct exchange. 98-100% exchange. P=15 kbar. Equation as in muscovite-calcite. Direct exchange. 44-90% exchange. P=15 kbar. Equation as in muscovite-calcite. Fluorophlogopite has F/(F+OH)=1 Direct exchange. 31-91% exchange. P=11 kbar. Equation as in muscovite-calcite. Fluorophlogopite has F/(F+OH)=0.75. Experimental data for this mixed fluoro-hydroxy phlogopite fall in between the equations for end-member hydroxy- and flurophlogopite-calcite fractionations given by Chacko et al. (1996). Experiments involved synthesis of muscovite from gels or reaction of paragonite or kaolinite with KCl solutions. Based on the experimental data of O’Neil and Taylor (1969), natural samples, and theoretical estimate of the effect of OH groups on fractionation behavior. Based on the experimental muscovite-water and quartz-water calibrations of O’Neil and Taylor (1969) and Matsuhisa et al. (1979) with a straight line constrained to go through the origin. Based on natural quartz-biotite data where temperatures were determined by the Bottinga and Javoy (1973) calibrations of the feldspar-quartz, feldspar-muscovite or feldspar-magnetite isotope thermometers.

Based on natural quartz-amphibole data where temperatures were determined by the Bottinga and Javoy (1973) calibrations of the feldspar-quartz, feldspar-muscovite or feldspar-magnetite isotope thermometers.

70 Chacko, Cole & Horita

Sheppard and Gilg (1996) Escande et al (1984)

Savin and Lee (1988)

Sheppard and Gilg (1996) Savin and Lee (1988) James and Baker (1976)

Sheppard and Gilg (1996) Bechtel and Hoernes (1990)

Wenner and Taylor (1971) Chiba et al. (1989) Zhang et al. (2001) Lowenstam(1962) O’Neil (1963)

kaolinite- H2O

smectite- H2O

smectite- H2O

smectite- H2O

illite- H2O

illite- H2O

illite- H2O

illite:framework-OH

serpentine- H2O Oxides and Hydroxides: magnetite-calcite magnetite- siderite

magnetite-H2O

magnetite-H2O

125

126

127

128

129

130

131

132

133

134 135

136

137

Ex

N

Ex Ex

N

N

Mx

Ex

Mx

Mx

Mx

Ex

Mx 25-95

0-350

0-350

3.31 (106/T2) - 4.82

2.58 (106/T2) - 4.19

2.55 (106/T2) - 4.05

800-1200 45-75

-5.91 (106/T2) -1.76 (106/T2) - 9.43

-5.5, -4.4

700, 800

9

150-400

1.56 (106/T2) - 4.70

5.6

200-300

0-350

2.39 (106/T2) - 3.76 -0.076 T(°C) + 30.42

22

23.4

2.39 (106/T2) - 4.19

0-350

2.76 (106/T2) - 6.75

Direct exchange. 82-100% exchange. P=15 kbar. Based on microbial siderite-water and magnetite-water fractionations established by thermophilic Fe(III)-reducing bacteria. Measurement of magnetite teeth from the species Cryptochiton stelleri, which grew in water at 9°C. Direct exchange. 84-94% exchange. Fractionations calculated by Matthews et al. (1983b) based on the experimental data of O’Neil (1963).

Regression line fit through a combination of experimental data (Kulla and Anderson, 1978) and various natural occurrences. Synthesis of Mg-rich smectite (stevensonite and saponite) under hydrothermal conditions. Savin and Lee (1988) point out that Escande et al.’s (1984) technique of analyzing smectites may have resulted in an overestimation of the smectite- H2O fractionation factor. Modification of the equation of Yeh and Savin (1977). Based on analysis of authigenic smecite formed at 1°C ('(sm-H2O)=30.3 ‰), natural smectite-illite pairs, the quartz-illite curve of Eslinger and Savin (1973), and the quartz-H2O curve of Matsuhisa et al. (1979). Regression line fit through a combination of experimental data (Kulla 1979) and various natural occurrences. Based on the natural sample quartz-illite curve of Eslinger and Savin (1973), and the quartz-H2O curve of Matsuhisa et al. (1979). Fractionation determined in experiments in which illite was suspended in 2N NaCl and 0.2N NaTPB-EDTA solution. The Northrop-Clayton extrapolation procedure indicates 140% exchange – a significant overshoot of the equilibrium fractionation. Thus, the derived fractionation factor may not be reliable. Regression line fit through a combination of experimental data (O’Neil and Taylor, 1969) and various natural occurrences. Determined intra-mineral fractionation by thermal dehydration and partial fluorination methods. Hamza and Epstein (1980) and Girard and Savin (1996) also attempted on many hydrous minerals, but the techniques are incomplete and the results are umbigouous. Same equation as that for chlorite- H2O.

Equilibrium Isotope Fractionation Factors 71

Ex Ex Ex

Mx Ex

Ex Ex

Fortier et al. (1995) Zhang et al. (1997) Mandernack et al . (1999) Downs et al. (1981)

Clayton and Epstein (1961) Yapp (1990) Bao and Koch (1999)

Müller (1995)

Bao and Koch (1999) Bao and Koch (1999) Chacko et al. (1996) Addy and Garlick (1974) Matthews et al. (1979) Bird et al. (1993) Bird et al. (1993) Matthews and Schliestedt (1984)

magnetite-H2O

magnetite-rich FeOH 2O magnetite-H2O

magnetite-quartz

hematite - H2O

hematite (geothite)H 2O

hematite - H2O

goethite - H2O

goethite - H2O

akaganeite (EFeOOH) - H2O rutile-calcite rutile-H2O rutile-H2O

rutile-H2O anatase-H2O rutile-quartz

140

141

142

143

144

145

146

147

148

149

150 151 152

153 154 155

Ex Ex Ex

Ex Ex Ex

Ex

Ex

Ex

Ex

Blattner et al. (1983)

magnetite-H2O

139

N

Bottinga and Javoy (1973)

magnetite-H2O

138

25-120 30-140

10-65

35-140 35-95 800-1000 575-775 300-700

1.63 (106/T2) - 12.3 0.733 (106/T2) - 6.914

1.10 (106/T2) - 12.1 (KOH) 0.3 (106/T2) - 3.0 (NaOH) 2.76 (106/T2) - 23.7 (hydrolysis) 1.907 (106/T2) - 8.004 3.927 (106/T2) - 12.157 -4.31 (106/T2) -4.1 (106/T2) + 0.96 -4.72 (106/T2) +1.62

22, 50 22, 50 500-700

25-120

0.413 (106/T2) - 2.56

6.1, 3.0 8.7, 4.9 -4.54 (106/T2)

600, 800

-7.8, -6.1

4-75

0.79 (106/T2) – 7.64

350 45-75

2

112, 175

0.80 (10 /T ) – 7.74

6

-8.60

-3.7, -7.9

-1.47 (106/T2) - 3.70

Direct exchange. 53-90% exchange. P=15 kbar. Exchange by crystallization of amorphous TiO2 powder. Exchange by controlled oxidation of Ti metal powder under hydrothermal conditions. Synthesis of rutile from TiCl4 solutions. Synthesis of anatase from TiCl4 solutions. Based on the experimental rutile-water and quartz-water calibrations of Matthews et al. (1979) and Matsuhisa et al. (1979).

Synthesis of geothite from FeCl3 solutions by the addition of NaOH solution. See above for details. Synthesis of akaganeite by hydrolysis of FeCl3 solutions.

Based on the analysis of feldspar-magnetite pairs from mafic lavas (Anderson et al.,1971), where solidification temperatures are relatively well known. Based on analysis of magnetite precipitated in steam pipelines of the Wairakei geothermal power station. Magnetite grown from fine-grained hematite. P = 1 kbar. Isotopic analysis by ion microprobe. Based on Fe3O4-rich iron oxides precipitated by Fe(III)reduction by thermophilic bacteria in culture. Based on microbial Fe3O4 grown within mangetotactic bacteria in culture and those of Zhang et al. (1997). Fayalite is oxidized to quartz and magnetite under hydrothermal conditions. P = 5 kbar. No equation given but regression line through the origin fitted to these two data points is in agreement with the results of Chiba et al. (1989). Based on the analysis of co-existing quartz, calcite and hematite and experimental calcite-water fractionation factor of Clayton (1961). Based on the synthesis of hematite (T• 62°C) or geothite (T95°C but differ significantly at lower temperatures. The authors attribute the discrepancy to differences in the washing and drying protocols applied to the hematite precipitates. Precipitation from Fe(NO3)2 soln by titrating KOH and NaOH solution, and by hydrolysis. The hydrolysis results differ significantly from the KOH and NaOH results.

72 Chacko, Cole & Horita

Ex

Xu and Zheng (1999) Fayek and Kyser (2000) Fayek and Kyser (2000) Zhang et al. (1994)

Lloyd (1968) Chiba et al. (1981) Kusakabe and Robinson (1977) Stoffregen et al. (1994) Stoffregen et al. (1994) Rye and Stoffregen (1995) Rye and Stoffregen (1995)

brucite-H2O

brucite-H2O uraninite(UO2)-H2O

UO3-H2O

cassiterite(SnO2)H 2O cassiterite(SnO2)H 2O scheelite(CaWO4)H 2O wolframite[(Fe,Mn) WO4]- H2O perovskite(CaTiO3)calcite Sulfates: anhydrite- H2O

anhydrite- H2O

barite- H2O

alunite(SO4)- H2O

alunite(OH)- H2O

jarosite (SO4)- H2O

jarosite(OH)- H2O

161

162 163

164

165

171

172

173

174

175

176

170

169

168

167

166

Ex

Bird et al. (1994) Vitali et al. (2000)

gibbsite- H2O gibbsite- H2O

159 160

Ex Ex Ex

Ustinov and Grinenko (1990) Zhang et al. (1994) Gautason et al. (1993)

Ex

Ex

Ex

Ex

Ex

Ex

Ex

Ex

Ex Ex

Ex

Ex Ex

N

Sushchevskaya et al. (1985)

Saccocia et al. (1998)

Chen et al. (1988)

gibbsite – H2O

158

N

Lawrence and Taylor (1971)

gibbsite – H2O

157

N

Agrinier (1991)

rutile-quartz

156

100-550 110-350 250-450 250-450 150-250 150-250

2.64 (106/T2) – 5.3 (salt-effect not corrected) 3.09 (106/T2) – 2.94 2.28 (106/T2) – 3.90 1.43 (106/T2) + 1.86 2.1 (106/T2) – 8.77

2

3.21 (10 /T ) – 4.72

6

100-500

200-420

25, 100

3.878 (106/T2) – 3.4

3

800-1000

2

25-450

250-370

100-300

15-120 100-300

3.13 (10 /T ) – 6.42(10 /T) – 0.12 -6.42 (106/T2)

6

8-51 0-60

1.31 (106/T2) – 1.78 2.04 (106/T2) – 3.61 (103/T) + 3.65 9.54 (106/T2) – 35.3 (103/T) + 26.58 1.56 (106/T2) – 14.1 16.58 (106/T2) – 77.52(103/T) + 77.48 -2.21 (106/T2) + 25.06(103/T) – 49.50 10.13 (106/T2) – 26.09(103/T) + 12.58 +14.5 (25°C) -1.1 to 0 (300-450°C) +10.8, +1.8 250-450

0-30

0-30

450-800

~+16

~+18

-4.78 (106/T2)

Direct exchange between anhydrite and water in 1N H2SO4 solution. P=690 bars. Direct exchange between anhydrite and water in 1m NaCl, HCl or H2SO4 solutions. 28-98% exchange. P=1-1000 bars. Direct exchange between barite and water in 1m NaCl or 1mNaCl-1m H2SO4 solution. 26-98% exchange. Cation exchange of natroanunite with 0.7m K2SO4. 8-95% exchange. Cation exchange of natroanunite with 0.7m K2SO4. 8-95% exchange. Cation exchange of natroanunite with H2SO4-K2SO4. 40100% exchange. Cation exchange of natroanunite with H2SO4-K2SO4. 40100% exchange.

Synthesis from Na2WO4, FeCl2, and MnCl2 in the presence of up to 30wt% NaCl or NaF. Direct exchange. 67-99% exchange. P= 15 kbar.

Precipitation from SnCl2 soln (25°C) and oxidation of Sn (300-450°C). Precipitation.

Synthesis by hydrolysis of Mg3N2 and MgCl2, and MgO. Combined experimental results of UO2-CO2 exchange with CO2 –H2O of Truesdell (1974). Combined experimental results of UO3-CO2 exchange with CO2 –H2O of Truesdell (1974). Synthesis from amorphous SnO2 or SnCl2 soln.

Direct exchange with 3.2 and 10wt% NaCl soln. P=500 bar.

Based on the analysis of quartz, rutile pairs in metamorphic rocks, primarily eclogites. Temperatures based on the Bottinga and Javoy (1975) calibrations of the quartzmuscovite and quartz-garnet thermometers. From analysis of samples from a large number of natural weathering profiles. Based on the analysis of gibbsite from bauxite deposits in Taiwan. By synthesis and aging for 3-56 months. By synthesis and aging for 3-56 months.

Equilibrium Isotope Fractionation Factors 73

Stolper and Epstein (1991)

CO2Silicate/Melt/Glass: CO2-silica glass

Feng and Savin (1993) Noto and Kusakabe (1997) Noto and Kusakabe (1997) Nahr et al. (1998)

Ex

Karlsson and Clayton (1990)

Ex

N

Ex Ex Ex

Mx

Karlsson and Clayton (1990)

analcime(channelwater)- H2O stilbite- H2O(v) wairakite- H2O wairakite (channelwater)- H2O clinoptilolite-water

Zeolites: analcime- H2O

220-300 250-400 250-400

-2.4 + 2.7 (106/T2) 2.46 (106/T2) - 1.76 0.79 (106/T2) - 3.07

+2.14 to +4.30

550-950

25, 40

300-400

1.01 (106/T2) - 8.87

+31.6±0.2, +26.6

25-400

2.78 (106/T2) - 2.89

Based on analysis of authigenic clinoptilolite samples in oceanic sediments from three ODP sites.

Direct exchange at low pressures (21 Torr). Direct exchange at 0.5-1.5 kbar. 63-98% exchange. Direct exchange at 0.5-1.5 kbar. 94-99% exchange.

Combined results of direct exchange at 300-400°C and two natural samples. Direct exchange at 1.5-5.0 kbar.

Direct exchange between silica wool or power and CO2 at 0.5 bar. Matthews et al. (1994) Ex +2.10 to +3.11 750-950 Direct exchange between silica wool or power and CO2 at 1 184 CO2-silica glass bar. Matthews et al. (1994) Ex +3.58 to +4.75 750-950 Direct exchange between albite glass and CO2 at 1 bar. 185 CO2-albite glass Matthews et al. (1994) Ex +3.36 to +4.74 750-950 Direct exchange between crystalline albite and CO2 at 1 bar. 186 CO2-albite Palin et al. (1996) Ex +2.95 to +5.08 550-950 Direct exchange between rhyolitic glass and CO2 at 0.8-1.5 187 CO2-rhyolitic glass/melt bar. Notes: 1. Ex = experimental; N = natural sample; M = mixed experimental/natural sample calibration. 2. Equations given with temperature in Kelvin unless otherwise indicated.

183

182

179 180 181

178

177

74 Chacko, Cole & Horita

Deuser and Degens (1967), Wendt (1968), Emrich et al. (1970), Mook et al. (1974), Turner (1982), Lesniak and Sakai (1989), and Zhang et al. (1995) Szaran (1997) Malinin et al. (1967) Turner (1982), Lesniak and Sakai (1989), Zhang et al. (1995) and Halas et al. (1997)

HCO3-(aq)-CO2(g)

HCO3-(aq)-CO2(g)

HCO3-(aq)-CO2(g) CO32-(aq)-CO2(g)

4

5

6 7

19

18

17

16

Ex

Ex

Melchiorre et al. (1999) Bottcher (2000)

Ex

Ex

Ex

Ex Ex

Ex Ex Ex Ex Ex

Ex Ex

Ex

Ex

Ex

Ex

Ex

Method

Bottcher (1994)

Matsuo et al. (1972)

Matsuo et al. (1972)

Romanek et al. (1992) Carothers et al. (1988)

aragonite-CO2(g) siderite (FeCO3) HCO3-(aq) gaylussite (Na2CO3·CaCO3· 5H2O)- CO32-(aq) trona (Na2CO3·NaHCO3· 2H2O)- CO32-(aq) pirssonite (Na2·Ca(CO3)2.· 2H2O)- CO32-(aq) CO2-malachite (CuCO3·Cu(OH)2) norsethite (BaMg(CO3)2) - CO2

13 14

15

Rubinson and Clayton (1969) Emrich et al. (1970) Turner (1982) Romanek et al. (1992) Rubinson and Clayton (1969)

calcite-HCO3-(aq) calcite-CO2(g) calcite-HCO3-(aq) calcite-CO2(g) aragonite-HCO3-(aq)

8 9 10 11 12

CarbonatesAqueous Solution:

Wendt (1968), Vogel et al. (1970), and Zhang et al. (1995)

CO2(aq)-CO2(g)

3

Grootes et al. (1969)

CO2:liquid-vapor

1

2

Eiler et al. (2000)

Reference

CO2:solid-vapor

CO2and Aqueous Species:

Phases

1

#

-0.57 to –0.14

-0.2 to +0.4

1000 ln D

60-90 0-50 20-90

-1.85 (106/T2 )+ 10.51 1.78 (106/T2) - 10.16

18-35

8-35

10-40 33-197

25 20-60 25 10-40 25

23-286 4-80

7-70

5-60

0-60

220-303(K)

130-150(K)

T(°C)

2.3 and 3.3

0.7 to 1.8

1.9 to 3.1

13.88 - 0.13 T(°C) -4.20 to +2.86

0.9±0.2 9.61±0.28 at 25°C 1.83 to 2.26 11.98 - 0.12 T(°C) 2.7±0.2

-7.3 to +7.5 -1.5 to +8.5

-0.0954 T(°C) + 10.41

7.92 to 8.27 at 25°C with –0.064 to –0.141‰/°C gradient

0.0041 T(°C)-1.18 (Vogel) 0.0049 T(°C)-1.31 (Zhang)

2

Formed from BaCO3 and MgCO3·3H2O in NaHCO3 soln.

Replacement of calcite with Cu2+-bearing soln.

Reaction of CaCO3 with Na2CO3 soln.

Formed from Na2CO3 and NaHCO3 solns onto seed crystal.

Formed by mixing Na2CO3 and CaCl2 solns.

Controlled seeded precipitation at constant composition. Slow injection of FeCl2 soln to NaHCO3 soln.

Slow precipitation by CO2 degassing. Slow precipitation by CO2 degassing. Slow precipitation by adding NaOH to Ca-HCO3 soln. Controlled seeded precipitation at constant composition. Slow precipitation by CO2 degassing.

Cross-over at about 150°C. Cross-over at about 67°C (Halas et al., 1997).

Also investigated the relaxation time.

Consistent results by many investigators.

C is depleted and 18O is enriched in the aqueous phase.

13

Liquid-vapor equilibration for >4 hr.

By sublimation-condensation and isotopic exchange.

Comments

Appendix 3. Carbon isotope fractionation factors: calibrations based on experiments or natural samples.

Equilibrium Isotope Fractionation Factors 75

Rosenbaum (1994) Scheele and Hoefs (1992) Scheele and Hoefs (1992)

Sommer and Rye (1978)

Hoering (1961) Javoy et al. (1978) Mattey et al. (1990) and Mattey (1991)

CO2-calcite CO2-calcite

CO2-graphite

CarbonatesGraphite/Diamond aragonite-calcite

dolomite-calcite

calcite-graphite

calcite-graphite

calcite-graphite

calcite-graphite

calcite-graphite

dolomite-graphite

diamond-graphite CO2 in Melts:

CO2-CO32- in melt

CO2-CO32- in melt

Gases: CO2-CH4

20

21 22

23

24

25

26

27

28

29

30

31

32

33

34

35

Ex

Ex

Ex

Ex

N

N

N

N

N

N

N

N

Ex

Ex Ex

Ex

270-650 400-800

8.9 (106/T2) - 7.1 5.81 (106/T2) - 2.61

4.0 to 4.6 1.8 to 2.7

1200-1400

1120-1280

>1700

400-680

5.9 (106/T2) - 1.9 0.3

700-800

3.56 (10 /T )

2

400-680

5.6 (106/T2) - 2.4

6

610-760

100-650

0.18 (106/T2) + 0.17 -0.00748 T(°C) + 8.68

0-25

600-1200

900 500-1200

400-950

2.56 - 0.065T(°C)

2.70 -3.46 (106/T2) + 9.58 (103/T) - 2.72 4.53 (106/T2) + 3.04

2.91 to 5.08 -0.10028 + 5.4173x - 2.5076x2 + 0.47193x3 – 0.049501x4 + 0.0027046x5 – 0.000059409x6 where x=106/T2

Solution of CO2 into silicate (sodamelilite and basalt) and carbonate melts at P=5-39 kbar.

Solution of CO2 into tholeiitic silicate melt at 7.0-8.4 kbar.

Based on analysis of coexisting calcite and aragonite tests and shells. Based on analyses of coexisting pairs of metamorphic origin. Preliminary experimental results show much larger fractionations. Based on analyses of metamorphic samples from the Adirondack Mountains. Based on analyses of metamorphic aureole samples from central Japan. Based on analyses of metamorphic aureole samples from central Japan. Based on analyses of metamorphic samples from Tudor aureole, Canada. Based on analyses of metamorphic samples from the Adirondack Mountains. Based on analyses of metamorphic aureole samples from central Japan. Conversion of graphite to diamond at P=70 kbar.

Direct exchange. 19-98% exchange. P=10 kbar. Large errors due to small degrees of exchange. Equation represents theoretical calculations that fit the experimental data of Chacko et al. (1991) and Rosenbaum (1994) within error. The data of Scheele and Hoefs (1992) are displaced to slightly larger fractionations. The equation reproduces the calculated fractionations from 273-4000K. Direct exchange. P=12.5 kbar. Most experiments involved inversion of aragonite to calcite. P=1-5 kbar. Direct exchange. 33-79% exchange. P=5-15 kbar. Small CO2/C ratio (1/32) used in the experiments may have resulted in the incorporation of surface effects in the measured fractionations.

26.70 - 49.137 (103/T) + 200-600 Complete exchange catalyzed by Ni catalyst. 40.828 (106/T2) 7.512(109/T3) Notes: 1. Ex = experimental; N = natural sample; M = mixed experimental/natural sample calibration. 2. Equations given with temperature in Kelvin unless otherwise indicated. Horita (2001)

Wada and Suzuki (1983)

Kitchen and Valley (1995)

Dunn and Valley (1992)

Morikiyo (1984)

Wada and Suzuki (1983)

Valley and O’Neil (1981)

Sheppard and Schwarcz (1970)

Chacko et al. (1991)

CO2Calcite/Graphite: CO2-calcite

76 Chacko, Cole & Horita

23

22

21

20

19

18

17

16

15

mirabilite (Na2SO4. 10H2O)- H2O

Hydrated Salts: trona (Na2CO3.NaHCO3. 2H2O) – H2O gaylussite (Na2CO3.CaCO3. 5H2O) – H2O natron (Na2CO3. 10H2O) – H2O borax (Na2B4O7. 10H2O) – H2O carnallite(KMgCl3. 6H2O) – H2O carnallite(KMgCl3. 6H2O) – H2O bischofite(MgCl2. 6H2O) – H2O tachyhydrite(CaMg2 Cl6. 12H2O) – H2O gypsum (CaSO4. 2H2O)- H2O Ex

Fontes and Gonfiantini (1967), Sofer (1978), and Matsubaya and Sakai (1973) Stewart (1974) Pradhananga and Matsuo (1985b)

Ex

Ex

Ex

Horita (1989) Horita (1989)

Ex

Koehler and Kyser (1996)

Ex

Ex

Ex

Pradhananga and Matsuo (1985b) Matsuo et al. (1972) Pradhananga and Matsuo (1985a) Horita (1989)

Ex

Ex

Ex Ex Ex Ex

Ex

Ex Ex Ex Ex Ex Ex Ex Ex

Matsuo et al. (1972)

Matsuo et al. (1972)

Horibe and Craig (1995) Suess (1949) Cerrai et al. (1954) Rolston et al. (1976)

CH4-H2 H2O(v)-H2 H2O(v)-H2 H2O(liq)-H2

10 11 12 13

14

Horita and Wesolowski (1994), and all references therein

H2O: liquid-vapor

9

Merlivat and Nief (1967) O’Neil (1968) Craig and Hom (1968) Arnason (1969) Suzuki and Kimura (1973) Stewart (1974) Lehmann and Siegenthaler (1991) Majoube (1971b)

+13 to +19

-15 to -20

-46 to -36

0-25

17-57

25-40

10-40

22-45

-18.4 (106/T2) + 162 -47 to -37

8-35 5-25 10-40

5-10

8-35

8-35

200-500 80-200 51-742 7-97

0-374

-40 – 0 0? 0? 0 0? -10 0 0-100

0 to +2 +2 to +5 -44 to -40

+15 to +17

-15 to -13

1.420 (107/T2) + 2.356 (104/T)

logD=-0.041 + 7.074/T2 +18.7±0.9 +19.5, +20.3 +20.6±0.7 +20.4±0.5 +24 +21.2±0.4 24.844 (106/T2) – 76.248(103/T) + 52.612 1158.8 (T3/109) - 1620.1 (T2/106) + 794.84 (T/103) + 2.9992 (109/T3) – 161.04 D=0.8994 + 183540/T2 logD=203/T-0.132 3.04 to 1.18 (D) ln D = -0.2143 + 368.9/T + 27870/T2

Synthesis from aqueous solutions. Determined on the composition scale of brine.

By precipitation and aging 30-67 days. Determined on the activity scale of brines. By precipitation and aging 30-67 days. Determined on the activity scale of brines. Slow precipitation.

Synthesis from aqueous solutions. Determined on the composition scale of brine. Synthesis from aqueous solutions. Determined on the composition scale of brine. By precipitation and aging 30-67 days. Determined on the activity scale of brines. By synthesis and exchange.

Synthesis from aqueous solutions. Determined on the composition scale of brine.

Synthesis from aqueous solutions. Determined on the composition scale of brine.

Dynamic exchange in a reactor with Pt catalyst. Exchange catalyzed by Pt catalyst.

High precision experimental calibration using three different apparatus to cover the full temperature range. This large dataset was combined with selected earlier experimental studies to generate the calibration curve. Complete exchange with Ni-thoria catalyst.

Freezing from 2.5m NaCl soln. Slow growth of ice in stirred water. Slow distillation of liquid water.

Slow growth of ice in stirred water. Freezing from 4‰ chlorinity water and seawater. Seeded-growth.

Appendix 4. Hydrogen isotope fractionation factors: calibrations based on experiments or natural samples. 1 2 Reference Method 1000 ln D T(°C) Comments

Gases, Liquid, Ice H2O: ice-vapor H2O: ice-liquid H2O: ice-liquid H2O: ice-liquid H2O: ice-liquid H2O: ice-liquid H2O: ice-liquid H2O: liquid-vapor

Phases

1 2 3 4 5 6 7 8

#

Equilibrium Isotope Fractionation Factors 77

Ex Ex

Ex Ex

Graham et al. (1980) Graham et al. (1980) Lawrence and Taylor (1971) Chen et al. (1988) Vitali et al. (2001) Vitali et al. (2001) Yapp and Pedley (1985)

Yapp and Pedley (1985) Yapp (1987) Hariya and Tsutsumi (1981) Satake and Matsuo (1984) Saccocia et al. (1998) Xu and Zheng (1999) Horita et al. (1999) Graham et al. (1980)

Chacko et al. (1999) Vennemann and O’Neil (1996)

Graham et al. (1980) Graham et al. (1980)

boehmite- H2O

diaspore- H2O gibbsite- H2O

gibbsite- H2O

gibbsite- H2O gibbsite- H2O

goethite- H2O(v)

goethite- H2O(l)

goethite- H2O(l) manganite- H2O(l) brucite-H2O

brucite-H2O brucite-H2O

brucite-H2O Sorosilicates: epidote- H2O

epidote- H2O

epidote- H2

zoisite- H2O

clinozoisite- H2O

26

27 28

29

30 31

32

33

34 35 36

37 38

39

41

42

43

44

40

Ex

Ex

Ex Ex

Ex Ex Ex

N

Ex

Ex N

N

Ex N

Ex

Ex

Ex

Suzuoki and Epstein (1976)

25

Pradhananga and Matsuo (1985b)

episomite (MgSO4. 7H2O)- H2O Hydroxides: boehmite- H2O

24

300-600 150-400

280-650

9.3 (106/T2) – 61.9 110.756 (106/T2) + 149.98 (103/T) – 158.685

-15.07 (106/T2) – 27.73

300, 450

200-300 300-650

29.2 (106/T2) – 138.8 -35.9

-37.8, -37.2

380

250-450 25-90

25, 62 150-250 144-510

~25

100, 145

9-60 Ambient

0-30

150, 200, 280, 380 380 0-30

400

25

-31.9 to –19.5

~-95 -211 to -192 8.72 (106/T2) – 3.86 (104/T) + 14.5 -32 to -22 4.88 (106/T2) – 22.54

-105

-75.8, -89.9

-2±6 -8±10

-8

-78.7 -15

-37.4, -36.2, -53.5, -49.9

-66.2

-1

Direct exchange. 30-100% exchange. P=2-4 kbar. Epidote composition is pistacite (Fe/Fe+Al) = 0.29. Experiments indicate fractionation factor is independent of temperature in the 300-650°C temperature range. Direct exchange. Experiments done on large single crystals and analyzed with the ion microprobe. P = 1.2 – 2.2 kbar. Epidote composition is pistacite (Fe/Fe+Al) = 0.33. Direct exchange of epidote with H2 gas. 71-91% exchange. P = 0.3-2 bars. At 300°C, epidote-H2O fractionations derived from these results are approximately 35‰ more negative than those reported in the direct epidote-H2O exchange experiments of Graham et al. (1980) and Chacko et al. (1980). Epidote composition is pistacite (Fe/Fe+Al) = 0.30. Direct exchange. 35-100% exchange. P= 4 kbar. Zoisite composition is pistacite (Fe/Fe+Al) = 0.03. Direct exchange. 60-100% exchange. P= 2 kbar. Zoisite composition is pistacite (Fe/Fe+Al) = 0.09.

Direct exchange with 3.2 and 10wt% NaCl soln. P=500 bar. Synthesis by hydrolysis of Mg3N2 and MgCl2. Large discrepancy with the data of Satake and Matsuo (1984). Increase with pressure from 150 to 8000 bar.

Direct exchange. 25% exchange. P = 4 kbar. From analysis of samples from a large number of natural weathering profiles. From analysis of samples from a large number of natural weathering profiles in Taiwan. Synthesis over 10 years. Based on analysis of 12 samples from various localities in the tropics. Direct exchange between natural goethite and water vapor. 52-75% exchange. P=1 bar. Corresponding ٛ oethite-H2O(l) fractionations are –103.6 at both 100 and 145°C using the liquid-vapor fractionation curve of Horita et al. (1994). Based on the analysis of natural goethite and associated modern waters. Synthesis of goethite from Fe(NO3)3 solutions. Direct exchange at 500 bar. 46-90% exchange. Direct exchange. 58-100% exchange. P=1-1060 bars.

Direct exchange. 93% exchange. P=1 kbar (Suzuoki, pers. commun.). Direct exchange. 77-100% exchange. P = 2 kbar.

Synthesis from aqueous solutions. Determined on the composition scale of brine.

78 Chacko, Cole & Horita

Ex Ex

Ex Ex Ex Ex

Graham et al. (1984) Suzuoki and Epstein (1976) Vennemann and O’Neil (1996) Suzuoki and Epstein (1976) Suzuoki and Epstein (1976) Suzuoki and Epstein (1976) Suzuoki and Epstein (1976) Vennemann and O’Neil (1996)

actinolite- H2O Micas: muscovite- H2O

muscovite- H2

lepidolite- H2O

phlogopite – H2O

biotite (Mg-rich) – H 2O biotite (Fe-rich) – H 2O biotite- H2

54

55

56

57

58

59

61

60

Ex

Graham et al. (1984)

tremolite- H2O

53

Ex

Ex

Ex

Vennemann and O’Neil (1996)

hornblende- H2

52

Ex Ex

Ex

Ex Ex N

Ex

Graham et al. (1984) Graham et al. (1984)

Suzuoki and Epstein (1976)

Amphiboles: hornblende- H2O

49

pargasite- H2O ferroan pargasitic hornblende – H2O

Blamart et al. (1989) Jibao and Yaqian (1997) Kotzer et al. (1993)

Cyclosilicates: tourmaline- H2O tourmaline- H2O tourmaline- H2O

46 47 48

50 51

Yaqian and Jibao (1993)

ilvaite – H2O

45

417.3, 293.0

300, 400

650

400-850

-21.3 (106/T2) – 2.8 -38.0

575, 650

650 -14.9, -9.9

316.851 (10 /T ) – 584.796 (103/T) + 515.684 +15.4

200-400

450-750

2

-22.1 (106/T2) + 19.1 6

400

650-850 350-650

-31.0 (106/T2) +14.9 -21.7 -29

300, 400

700, 850 850-950 350-700

400-700

500-600 450-800 350-600

350-750

452, 326

-21.86, -18.12 -31.0 (106/T2) + 1.1 -23.1

-23.9 (106/T2) + 7.9

-39.3 (106/T2) + 63.4 -27.9 (106/T2) + 2.3 -27.2 (106/T2) + 28.1

-105 (350-550°C) -29.95 (106/T2) – 60.62 (550-750°C)

Direct exchange. 39-100% exchange. P=1 kbar (Suzuoki, pers. commun.). Direct exchange of muscovite with H2 gas. 13-80% exchange. P =0.4-2 bars. Direct exchange. 45% exchange. P=1 kbar (Suzuoki, pers. commun.). Direct exchange. 36-81% exchange. P=1 kbar (Suzuoki, pers. commun.). Biotite composition is Mg/(Fe+Mg) = 0.89. Direct exchange. 31-100% exchange. P=1 kbar (Suzuoki, pers. commun.). Biotite composition is Mg/(Fe+Mg) = 0.61. Direct exchange. 100% exchange. P=1 kbar (Suzuoki, pers. commun.). Biotite composition is Mg/(Fe+Mg) = 0.33. Direct exchange of biotite with H2 gas. 31-64% exchange. P =2 bars. Extensive reduction of Fe3+ and increase in water accompanied isotopic exchange.

Direct exchange. 19-100% exchange. Pressure unspecified. Composition of the hornblende not clearly specified but appears to actinolite or actinolitic hornblende. Direct exchange. 83-90% exchange. P = 4-6 kbar. Direct exchange. 46-100% exchange. P = 2-8 kbar. Experiments indicate fractionation factor is independent of temperature in the 350-700°C temperature range. Direct exchange of hornblende with H2 gas. 71-91% exchange. P = 2 bars. Fe3+ content of the hornblende decreased and water content increased during the experiment indicating that isotopic exchange was accompanied by partial reduction of Fe. Direct exchange. 75-99% exchange. P = 2-4 kbar. Experiments indicate fractionation factor is independent of temperature in the 350-650°C temperature range. Direct exchange. 59% exchange. P = 2 kbar.

Direct exchange at P=3 kbar. 72-92% exchange. Direct exchange at P=150-250 bar. 11-81% exchange. Based on the analysis of co-existing quartz, muscovite and tourmaline from several ore deposits. Temperatures based on quartz-muscovite oxygen isotope fractionations compiled by Eslinger et al. (1979).

Direct exchange at P=50-203 bar. 50-100% exchange.

Equilibrium Isotope Fractionation Factors 79

Mx

Ex

Wenner and Taylor (1973) Suzuoki and Epstein (1976) Sakai and Tsutsumi (1978) Grinenko et al (1987) Mineev and Grinenko (1996)

Suzuoki and Epstein (1976) Lambert and Epstein (1980) Liu and Epstein (1984) Gilg and Sheppard (1996)

Vennemann and O’Neil (1996) Yeh (1980)

Capuano (1992)

serpentine- H2O

serpentine- H2O

serpentine- H2O

serpentine- H2O

kaolinite-H2O

kaolinite-H2O

kaolinite-H2O(v)

kaolinite-H2O

kaolinite- H2

smectite-H2O

illite/smectite-H2O

63 64

65

66

67

68

69

70

71

72

73

74

75

N

N

Ex

N

Ex

Ex

Ex

Ex

Mx

N Ex

Taylor (1974) Graham et al. (1987)

chlorite- H2O chlorite- H2O

62

N

Marumo et al. (1980)

Other Hydrous Phyllosilicates: chlorite- H2O

100-500

2.75 (107/T2) - 7.69 (104/T) + 40.8 -40 to +35 (100°C) -23 to +6 (200°C)

250-352 0-330

150-275 25-120

0-150

-2.2 (106/T2) - 7.7

-162.495 (106/T2)+ 1241.164 (103/T) – 1219.110 -19.6 (103/T) + 25

-45.3 (103/T) + 94.7

strontianite (Sr) > cerrusite (Pb) > rhodocrosite (Mn) > calcite (Ca) ≥ dolomite (Ca, Mg) > magnesite (Mg). This order follows approximately the magnitude in the rc + ro distances, from greatest to smallest. The order for the layer silicates, from fastest to slowest, is chlorite > biotite > phlogopite > margarite > paragonite ≥ muscovite (Fig. 17b). Qualitatively, this order follows a pattern of increasing Si, Al, Na, K and Ca and decreasing Mg and Fe. The order of oxygen isotope rates for these layer silicates parallels the order for decreased dissolution rates described by Brady and Walther (1989) that is

126

Cole & Chakraborty

used to explain the inverse mineral reaction series sequence of surface weathering. By establishing an unambiguous relationship between rate, lattice energy, and ultimately temperature, we can begin to develop empirical equations useful in predicting rates of isotopic exchange for minerals for which experimental data are lacking. DIFFUSION-CONTROLLED MINERAL-FLUID ISOTOPE EXCHANGE

The rate of isotopic exchange under many geological conditions is dependent on the type and rate of diffusion of the isotopic-bearing species in the system (e.g. CO2, H2O). The study of diffusion of isotopic-bearing species is a subset of the much broader topic of diffusion-controlled transport of elements in mineral-fluid, mineral-gas, mineral-mineral, mineral-melt and melt-volatile systems. Indeed, it is now recognized that intercrystalline diffusion is fast enough to play a major role in controlling metamorphic and hydrothermal reactions. Diffusion can lead to profound chemical and isotopic heterogeneities whose distributions can be used to decipher the thermal histories of individual mineral assemblages (Valley, this volume). Of interest is whether or not the mineral grains have maintained their mutual chemical (and isotopic) equilibrium by diffusion and when, during cooling of the rock, the mineral compositions became frozen in (Dodson 1973). This issue of chemical and isotopic compositions continuing to readjust during cooling below a peak thermal event—i.e. closure temperature (and pressure)—complicates the interpretation of the thermal history of any rock system, but, if properly constrained, can also provide more quantitative insight into the pressuretemperature-time paths. However, to do this requires appropriate diffusion data and an understanding of the mechanisms controlling diffusion. The purpose of this section is to (a) describe the fundamental concepts important for understanding isotopic exchange controlled by diffusion in crystalline phases, (b) outline the methods used to determine diffusivities, and (c) document in some detail the factors that control diffusion. Let us consider the definition of diffusion in general terms as the random jump of particles of matter (atoms, ions, molecules) from one site to another in the medium. This definition clearly presumes the existence of sites—i.e. locations within the crystal (or melt), characterized by a certain geometry and local structure—where an atom of a specific isotope may prefer to reside (Chakraborty 1995). Preference in this case refers to locations of low potential energy where the atom or ion spends more time occupying or vibrating around these sites relative to the time of jump from one site to another, which is a fundamental difference between diffusion in solids from that in an aqueous solution or a gas, with silicate melts/glass providing a case of transition between the two kinds of behavior. A number of mechanisms have been proposed for the elementary atomic jump (Fig. 18, from Bocquet et al. 1996). Most involve one or more kinds of point defects, without which diffusion is impossible in a crystalline solid. For crystalline materials, the generally accepted mechanisms involve the motion of one of two types of point defects, vacancies caused by the absence of an atom or ion at a lattice site (monovacancies, divacancies, and higher order vacancy clusters) or interstitials where an additional atom, ion or molecule occupies a normally unoccupied site (single interstitials, dumbbell interstitials, interstitialcies, and crowdions) (Peterson 1975). These lattice defects which control lattice (or volume) diffusion, are generated either thermally (intrinsic; e.g. Schottky, Frenkel; see Lasaga 1981c) or by the presence of impurities (extrinsic). Identification of the individual atomistic mechanisms is, however, not straight forward. Isotope mass effect studies are one approach and these have been conducted primarily on oxides that compared diffusivities of 18O with 17O. A discussion of this technique will be described below in more detail. In addition to isotopic exchange occurring through a volume diffusion mechanism,

Rates and Mechanisms of Isotopic Exchange

127

Figure 18. Mechanisms of diffusion, modified from Bocquet et al. (1996): (1) direct exchange, (2) cyclic exchange, (3) vacancy, (4) interstitial, (5) interstitialcy, (6) crowdion.

faster short-circuit pathways may also play a role depending on the nature of the solid (Lasaga 1981c). Macroscopic defects having dimensions exceeding 10-3 cm can be present as cracks, fissures and empty pore space. On the microscopic scale, dislocations or other line defects may facilitate diffusion and can consist of displacements of lattice planes along a defect boundary line. Planar defects such as stacking faults, twin boundaries and low and high angle grain boundaries may also be present. As we will see later, grain boundary diffusion has been the focus of a number of recent studies where isotopically–enriched water was reacted with fine-grained polycrystalline aggregates (e.g. Faver and Yund 1995). Non-planar defects have also been identified and include pipes (Yund et al. 1981) and tubes (Hacker and Christie 1991). Sitzman et al. (2000) imaged fast pathways in magnetite by TEM which correlated with δ18O variations determined by the ion microprobe. Finally, surface diffusion can occur when free surfaces are exposed to fluids (liquids or gases), and under these conditions mineral-fluid exchange may also involve dissolution and precipitation (see e.g Robertson 1969, Steele and Kilner 1982, Grabke 1994, Martin and Duprez 1996, Doornkamp et al. 1998). Numerous surface diffusion studies have been conducted on dry systems where a gas was reacted with a solid, usually an oxide, in order to assess quantitatively the catalytic properties of the solid surface. Discussion of this topic is beyond the scope of this review and will not be considered further. The reader is encouraged to consult the complete texts devoted to diffusion theory and practice, including Jost (1960), Manning (1968), Crank (1975), Allnatt and Lidiard (1993), Lasaga (1998) and Glicksman (2000). Fick’s laws

To measure the motion of atoms (or any other particle) one defines the physical quantity flux (or more rigorously, flux density), J, which is a vector characterized by a magnitude and direction. A flux, Ji, gives the quantity of a diffusing atom (i), which passes per unit time through unit area of a plane perpendicular to the direction of diffusion. Representing the concentration, or amount per unit volume, as Ci , Fick’s first law may be written as

128

Cole & Chakraborty

Ji = -D∇Ci

(105)

where ∇ is the mathematical shorthand for gradient, ∂Ci/∂x in one dimension (e.g. tabular crystal) or ∂Ci/∂r for radial diffusion (e.g. spherical grain with radius r). This law is similar in form to Fourier’s law stating that the flow of heat is proportional to the temperature gradient. The factor D, known as the diffusivity or diffusion coefficient, was introduced as a proportionality factor. The negative sign is a matter of convention that ensures that for positive diffusion coefficients, flux is positive in the direction of decreasing concentration, i.e. net motion is down concentration. Because the gradient involved is a “local gradient” (in space and time), Fick’s first law is by no means a “steady state” or a “time-independent” law (in the sense that it is valid only at steady state), as some text books would like to imply (Chakraborty 1995). It should be noted that D has the dimensions of (area/time), where D is usually given as cm2/sec or m2/sec in the S.I. system. In an isotropic mineral, the diffusion coefficient is independent of direction and there is only one value for D. For an anisotropic mineral the diffusion coefficient may vary many orders of magnitude in different directions. Fick’s first law relates flux to concentration gradient at a given place and time, but it does not explicitly describe how the concentration at a point evolves with time. To obtain a complete description of the process as a function of both position and time, it is necessary to combine Equation (105) with the equation that describes the conservation of atoms of type i

∂Ci/∂t + ∇ · Ji = 0

(106)

Elimination of Ji between Equations (105) and (106) gives Fick’s second law for one dimension and

∂Ci/∂t = D ∇2Ci

(107)

∂Ci/∂t = D [∇2 + (2/r) ∇]

(108)

for radial diffusion in a sphere. The physical interpretation of these equations is quite 2 2 straightforward. The terms ∇ Ci or [∇ + (2/r) ∇] represent the curvature of the diffusion profile. Therefore, whenever the concentration profile (i.e. Ci versus x) is concave downwards in a given region, the concentration decreases with time. Solutions of this partial differential equation giving Ci as a function of time and position can be obtained by various means once the boundary conditions have been specified. In simple cases (e.g. where D is a constant and the conditions are highly symmetric) it is possible to obtain explicit analytic solutions of Equation (107); see e.g. Carslaw and Jaeger (1959). Experiments can often be designed to satisfy these boundary conditions, but in many other practical situations numerical methods must be used which accommodate different grain geometries (e.g. Crank 1975). In the context of motion and flux, it is clear that the flux should be defined with respect to a reference frame (Chakraborty 1995). In crystalline silicates, because diffusion of oxygen and silicon can be much slower than that of other cations (with possible exception of Al), this can be achieved quite easily by using the fixed silicate lattice as a reference frame in which ions jump from site to site. This is the so-called lattice fixed frame, which commonly coincides with the laboratory frame. Note that the motion of a dilute isotope of oxygen (e.g. 18O) can still be treated in this frame. Fick’s first law can readily be modified to take into account the variability of reference frames.

Rates and Mechanisms of Isotopic Exchange

129

It can be further extended to describe simultaneous diffusion of multiple species in a multicomponent solid. However, Fick’s law may not effectively handle individual atomic jumps at short spatial scales or on very short time scales (as determined by some spectroscopic methods or computer simulations), nor can it address cases where diffusion occurs in a medium whose structure is changing, like the glass transition region. These may lead to what is termed non-Fickian behavior (e.g. Crank 1975). Diffusion coefficients

Solutions to Fick’s equations for a variety of boundary conditions and grain geometries, i.e. infinite plate, cylinder, sphere, etc., have been derived by numerous workers (e.g. Jost 1960, Crank 1975). Many of these solutions demonstrate how the degree of equilibration, F (see e.g. Eqn. 97), is related to the diffusion coefficient, as well as grain size, grain geometry and solution to solid volume ratio. Diffusion rates are sensitive to a number of factors which can be broadly divided into (a) environmental and (b) crystal chemical (Freer 1980). The former includes temperature, pressure, water fugacity, and oxygen activity. The latter includes chemical purity, defect concentration, crystal orientation for anisotropic phases, and porosity of either the single crystal or polycrystalline aggregate. A number of different types of diffusion coefficients (D) are encountered in the literature and the type of diffusion must be known if the determined D is to be a meaningful value. In some commonly used equations to solve for D, it is assumed that D is a function of Ci, which does not change appreciably during the experiment (i.e. the composition of the solid is essentially the same after reaction). This is accomplished either (a) by using a measurement technique (commonly involving radioactive tracers) that can detect very small changes in Ci, or (b) by using diffusional exchange of stable isotopes of the same element that leave the element concentration unchanged (Brady 1995). Approach (a) yields a tracer diffusion coefficient, D*, which, according to some authors, refers to transport of atoms present only in dilute quantities. Approach (b) yields a self-diffusion coefficient for the isotopically doped element that is also specific to the bulk chemical composition. Some confusion exists regarding which of these two terms describe transport of an isotope of the host’s own species through itself, where the isotope is present at a trace concentration. Consequently, the two have been used interchangeably to some extent in the literature. Both approaches generally ignore the opposite or exchange flux that must occur in dominantly ionic phases such as silicate minerals, glasses and melts. Experiments that use isotopically labeled species interacted with a mineral or mineral aggregate (grain boundary diffusion) are by far the most prevalent with regard to the determination of diffusivities (Table 2, Appendix), and will be the focus of our remaining discussion. The other major category of D values are termed interdiffusion or chemical diffusion coefficients. Experimentally, Ci does change significantly and D is a function of Ci. Diffusion of one or more chemical species is dependent on the opposing diffusion of another species in order to maintain a constant matrix volume and/or electrical neutrality. The diffusion in olivine of Mg2+ in one direction and the complementary diffusion of Fe2+ in the opposite direction represent one example. Rarely does this type of experiment employ the use of isotopically labeled species. However, in some cases isotopicallyenriched H2O (T, D and/or 18O) has been used where the composition of the solid (melt) became significantly modified by incorporation of water into the structure. With some loss of accuracy, therefore, “tracer” and “self” diffusion coefficients may be used interchangeably to refer to diffusion of a species in the absence of a driving force, e.g. a chemical potential gradient in a non-ideal solution. Similarly, “chemical” and

130

Cole & Chakraborty

“inter-diffusion” coefficients may be used interchangeably to refer to diffusion of the same species in the presence of a driving force, e.g. a chemical potential gradient in a non-ideal solution (i.e. concentration gradients in most real, multicomponent systems). The important point to note is: the use of one or the other kind of diffusion coefficient to describe a process does not depend on the species concerned (element or specific isotope of an element), but rather on the nature of the boundary conditions of the specific diffusion problem, e.g. whether there is a concentration gradient present or not. This is in contradiction to the fairly common practice of using “tracer” or “self” diffusion coefficients wherever modeling transport of any isotope is concerned. Secondly, under specific circumstances, the “tracer” or “self” diffusion coefficients may be numerically very similar to the “chemical” or “inter”- diffusion coefficients. This is in particular the case when the diffusing species is very dilute and the diffusing medium is a nearly ideal solution with respect to mixing of the diffusing units. In these cases, all four diffusion coefficients can be used interchangeably for practical purposes, although their physical significance remains different in spite of the numerical similarity. Relationships between tracer and chemical diffusion coefficients in multicomponent amorphous silicates have been derived by a number of authors over the years; we refer the reader to Liang et al. (1997) who suggest their own model and then compare various models with experimental data to decide on the most appropriate one for silicate melts, and to Lasaga (1998) who elucidates the theoretical relationships between the different kinds of models. For crystalline silicate systems, the model of Lasaga (1979a) has been shown to be quite successful and for aqueous solutions, the model given by Lasaga (1979b) has been demonstrated to be almost exact (Applin and Lasaga 1984). A related question that arises automatically is: do isotopic gradients homogenize more rapidly than the corresponding elemental concentration gradients? Some experimental studies on silicate melts demonstrated faster homogenization of isotopic ratios relative to the corresponding elemental gradients (e.g. Lesher 1994, van der Laan et al. 1994), and this has led to a variety of discussions in the literature. In answering this question, it is useful to remember that the diffusive flux of a species depends on the diffusion coefficient as well as the concentration gradient, J = -D(∂C/∂x) [see below for a more detailed form of the flux equation in glasses/melts]. Thus, a smaller concentration gradient, e.g. that of one particular isotope of an element, may be homogenized faster than the overall gradient of the element concerned. This observation, however, does not imply that diffusion of that particular isotope is governed by a different diffusion coefficient (e.g. tracer diffusion coefficient)—rather, it is a simple consequence of the abundance of the specific isotope and the magnitude of the relevant concentration gradients (overall elemental as well as that of the specific isotope). Quantitative analyses to clarify this point for silicate melts may be found in Richter et al. (1999) or Ozawa and Nagasawa (2001); Lesher (1994) presents and discusses experimental data to address this issue. A corollary of this discussion is that both—chemical, as well as tracer diffusion coefficients—are of interest for modeling the evolution of stable isotopes in condensed systems. Incidentally, a consequence of the experimental (Richter et al. 1999) and numerical simulation (Tsuchiyama et al. 1994) work is that the isotopic fractionation during diffusion in melts is much lower (depending on a mass exponent of 0.1 or less, as opposed to ~0.5 in the gaseous state), such that detectable mass fractionation is unlikely to result from diffusion in molten systems. Determination of D

As noted above, in the case of isotopic exchange, we generally measure the selfdiffusion coefficient which is defined as the rate of movement of a given isotopic species through a host with a bulk composition that is composed, in large part, of that species

Rates and Mechanisms of Isotopic Exchange

131

(e.g. 18O in silicates, 13C in carbonates). A wide variety of experimental configurations have been used to determine D values, and these can be divided into essentially three groups: bulk techniques, single-crystal techniques, and mineral aggregates. The experimental boundary conditions are commonly dictated, in large part, by the choice of the analytical method available to characterize the change in isotopic composition, either as a function of time, or preferably, space. Bulk exchange methods typically use isotope ratio gas source mass spectrometry (IRMS), as well as NMR, FTIR, liquid scintillation, and even less sophisticated approaches such as weight gain (gravimetry). The singlecrystal and mineral aggregate studies rely heavily on secondary ion mass spectrometry (SIMS or the ion microprobe) and nuclear reaction analysis (NRA), and to a lesser extent Rutherford Backscattering (RBS), FTIR and Mössbauer spectroscopy. Grain boundary diffusion studies typically involve the use of mineral aggregates, which are analyzed with the ion microprobe. Interestingly, under certain circumstances single crystal studies can also be used to address fast diffusion pathways such as dislocations, as well as document multiple diffusion mechanisms within single crystals (e.g. Moore et al. 1998). It is important to note that most of the diffusion data summarized in Table 2 (see Appendix) were obtained in order to quantify either the transport rate of the element of interest (e.g. O, C) or the solid state properties of the crystalline phase, chiefly, the nature of defects. Because most of these studies used isotopically labeled compounds, we assume that the rates of isotopic exchange can be adequately represented by these diffusivities. Therefore, the utility of diffusion data in modeling natural systems depends on selection of the appropriate D and its quality. What constitutes a successful (ideal) diffusion experiment? 1. Precise and accurate measurement of concentration as function of distance (depth) and time, 2. measurement of D at various temperatures and pressures, water, oxygen and hydrogen fugacity, 3. control of the P-T-X conditions so that the solid remains unreacted with either the gas or fluid phase (buffering if necessary; use SEM and/or TEM to characterize the solid), 4. use of different chemical compositions within a particular mineral series (e.g olivine, feldspar) to assess role of cation chemistry, 5. the solid phase is chemically and isotopically homogeneous and the defect state is well-characterized (e.g. phase has been properly annealed prior to reaction), and for a synthesized phase, its formation conditions are well constrained, 6. for bulk exchange experiments, detailed knowledge of the grain size and shape before and after the run are a must, 7. selective use of doped impurities (at ~10 to 103 ppm level) can help constrain the defect migration rates that influence diffusion, 8. use of multiple isotopes (e.g. 18O and 17O) to quantify the correlation factor. Bulk exchange methods. Prior to about 1980 the majority of the diffusion coefficients reported in the literature were obtained from a bulk exchange technique. In this approach, fine-grained crystalline powders of silicate, oxide, carbonate, etc. are reacted with a volatile phase such as H2O, O2, CO2, or CO2+CO. In numerous examples, at ambient pressures (0.1 MPa), copious amounts of gas or water vapor (±H2) are passed over the mineral powder such that the isotopic composition of the volatile does not change appreciably during the course of the experiment. Therefore, only the isotopic change in the mineral need be determined. Jost (1960) described a method for calculating

132

Cole & Chakraborty

the diffusion coefficients from partial exchange experiments using a simple equation for the boundary conditions of an infinite reservoir of isotopically constant gas (H2O). For a spherical grain geometry, this expression can be cast in terms of isotopic compositions as

δ 18Of − δ 18 Oeq 6 n =∞ 1 2 = ∑ (exp[−n t / τ ) δ 18O i − δ 18Oeq π 2 n=1 n 2

(109a)

⎧⎪ π 2 ⎛ δ 18O − δ 18Oeq ⎞ ⎪⎫ ln ⎨ ⎜ 18 f 18 ⎟ ⎬ ≅ −t / τ ⎪⎩ 6 ⎝ δ Oi − δ Oeq ⎠ ⎪⎭

(109b)

where D is the diffusion coefficient (cm2/sec), t is time in sec, τ = (r2/π2D), r is the grain radius in cm, and the δ18O compositions are designated by i (the isotopic composition of the starting material), f (isotopic composition of the exchanged solid), and eq (the equilibrium composition of the solid). Diffusion coefficients can be calculated from the slope of a plot of degree of exchange (left hand side of Equation 109b) versus time for several partial exchange experiments (see e.g. Muehlenbachs and Kushiro 1974). Conventional gas source isotope ratio mass spectrometry is used to determine the isotopic compositions. The bulk method was used by Muehlenbachs and Kushiro (1974) and Connolly and Muehlenbachs (1988) to determine diffusivities in silicates reacted with isotopically normal CO2 at elevated temperatures. Some studies have used this same approach, but instead reacted small volumes of a volatile with a large quantity of solid and measured only the isotopic changes in the gas phase. Alternatively, 18O-enriched O2, CO2 or H2O have been reacted in small to modest quantities with solids so that the condition of an infinite 18O reservoir is still satisfied. In many of these cases, the solid is commonly pre-annealed (equilibrated) at the desired run temperature with a gas of normal 18O/16O ratio prior to the introduction of the isotopically enriched gas. A significant number of diffusion studies that utilized the bulk isotope exchange approach were conducted in systems where diffusion occurred from a well-mixed gas or fluid reservoir of limited volume. Analytical expressions have been derived by Crank (1975) that relate the extent of exchange with diffusivity as a function of grain geometry, grain radius (sphere or cylinder) or grain width (plate), and the volume ratio of volatile (solution) to solid corrected for equilibrium isotope partitioning (see Cole et al. 1983, Lasaga 1998). This approach is valid, however, only for experimental conditions where the solution to solid mass ratio does not exceed approximately 10 for oxygen isotope exchange, and about one for hydrogen isotope exchange. For small values of (Dt/r2)1/2 in a sphere (10

4.5

8.5-9.1

4-7

2.3-5.5 6.1-7.3 2.0-2.4

(9.0)

(7.0)

Catalyst or pH (b)

Rates and Mechanisms of Isotopic Exchange 193

194

Cole & Chakraborty Table 2. Summary of rate data from

No.

Solid

Temp

Pressure Range

Sample

Reacting

Isotope

(a)

Range

(Mpa)

Type

Species

(e)

(ºC)

(b)

(c)

(d)

I. Framework Silicates A. Silica Group 1

D Quartz

400-620

2.5

SNX

H2O

H-D

2

E Quartz

700-900

2.5

SNX

H2O

H-D

3

E Quartz

1010-1220

700

445 961 931 719 553 578 727 1439 1267 1011 1185 954 1757 767 1435 588 630 691 1207 973 903 976 1355 942 1008 1124 1287 >700

364 166 651

55 55 55

544.3 420.3 1367

618.1 541.3 1534

706.2 710.8 1738

813 964 1993

E KJ/mol

CARBON calcite calcite graphite

~isotropic ~isotropic ~isotropic

0.02-24 b dry

*Footnotes: Note that cooling rate scales inversely to radius squared in Eqn. (9). Thus, Tc for 106C/Ma and 0.1mm is equivalent to Tc for 104C/Ma and 1mm, 102C/Ma and 1cm, etc. See Cole and Chakraborty (2001) for a discussion of alternate calibrations of Do and E. Do and A estimated by the technique of Fortier and Giletti 1989 (Ghent and Valley 1998).

Aluminosilicate-quartz Assemblages of quartz plus kyanite, sillimanite or andalusite have yielded exceptionally precise and apparently accurate temperatures when applied as a RAM thermometer. Diffusion of oxygen in aluminosilicates is very slow, yielding TC > 800°C for moderate to coarse grain sizes, while diffusion is comparatively fast in quartz (Table

Stable Isotope Thermometry at High Temperatures

387

3). The effects of solid solution are generally nil and the temperature coefficient relative to quartz is moderate (Zheng 1993b, Sharp 1995, Tennie et al. 1998). Putlitz et al. (2001) analyzed coarse kyanite from deformed quartz veins in pelites from the kyanite (+sillimanite) zone on the island of Naxos, Greece (Table 4). Six samples from an outcrop test within 100m at Stavros yielded highly precise fractionations (±0.06 ‰) and temperatures in excellent agreement with published estimates for coexisting kyanite + sillimanite. Temperatures above TC(Qt) are preserved because there is not another low Tc mineral in the veins. The host pelitic rocks at Stavros also contain quartz and kyanite; however, Δ(Qt-Ky) is larger in pelites (3.03 vs. 2.62 ‰) due to diffusive exchange between quartz, feldspars, and micas yielding reset temperatures that are 65°C lower. Fibrolitic sillimanite was also analyzed from three samples at Stavros, yielding about the same average temperature, but more variable results (634-726°C), possibly reflecting exchange due to the fine grain size. Samples of Qt-Ky from other localities accurately reproduce the petrologic temperatures. These accurate and precise results demonstrate the potential of RAM thermometry. Table 4. Aluminosilicate- quartz thermometry in quartzites, pelites, and quartz veins. Average Qt-AS Independent Ref. 1.

2.

Location

Δ(Qt-AS)

n

Ave T*

Naxos, Greece, deformed quartz veins, coarse kyanite Stavros 2.62±0.06 6 659±11 Komiaki 2.48 1 685 Sifones 2.64 1 656 Moni 2.65 1 649 Appollon 2.76 1 635 Appollon Village 2.77 1 634 Mica Creek, BC, pelites and quartz nodules Qt-Ky, nodule 2.55 2 665±10 Qt-Ky, pelite 3.0±0.24 4 596±35 Qt-Ky, pelite 2.9±0.25 3 605±38 Qt-Si, pelite 2.7±0.25 3 648±43

T estimate 660 670 630 630 630 620 645 637±28 605±26 702±23

Ky-Si zone St-Ky zone Ky-Si zone Si zone

References: 1. Putlitz et al. (2001); 2. Ghent and Valley (1998). * T calculated from Sharp (1995).

In an experimental study, Tennie et al. (1998) challenge the interpretations of Sharp (1995) and Ghent and Valley (1998) for aluminosilicate-quartz thermometry. They propose that the refractory nature of kyanite prevents metamorphic equilibration and makes kyanite generally unfit for thermometry. They estimated AQt-Ky = 3.00 (Eqn. 4) based on 11 of 17 piston cylinder experiments for calcite-kyanite exchange, and analysis using externally heated nickel reaction vessels. This calibration yields higher temperatures than the A value empirically estimated by Sharp (A = 2.25, 1995). In many instances, temperatures based on A = 3.00 are in significant disagreement with independent estimates. For instance, at Stavros (Naxos, Table 4), the average temperature is raised from 659 to 797°C. At 797°C, there should be widespread melting in these water-rich metasediments, but no evidence of in situ melting is observed on Naxos except at much higher grade. Tennie et al. ascribe such discrepancies to slow diffusion in kyanite (see Table 3) which could prevent exchange and cause kyanite to preserve δ18O from the lower temperatures of first crystallization rather than the peak of metamorphism. However, this explanation would yield Qt-Ky temperatures in quartzite that are too low rather than too high as is observed, and it does not explain the excellent agreement of

388

Valley

results from six different hand samples in the outcrop test at Stavros (Table 4). Discrepancies between empirical and experimental calibrations exist for several systems, including quartz-alumino-silicate. In this case, one should ask if an empirically derived thermometer falsely seems to record accurate temperature because the rocks used for calibration were retrogressed by an equal amount as those being studied, or if all rocks really preserve equilibrium compositions. This distinction is important because a non-equilibrated system cannot be relied upon and unequilibrated apparent temperatures may actually be controlled by other variables such as fluid composition, time, or deformation. Conversely, if careful thermometry yields self-consistent results in agreement with other systems, as concluded by Sharp (1995), Ghent and Valley (1998), Vannay et al. (1999), and Putlitz et al. (2001), then it may be that the experiments are in error. Values of Δ(Qt-Al2SiO5) can yield more than temperature information. Larson and Sharp (2000) analyzed coexisting quartz + sillimanite + kyanite + andalusite to show that “triple-point” assemblages from New Mexico did not form in equilibrium. Texturally equilibrated quartz + andalusite + sillimanite from the Front Range, Colo, yield estimates of pressure as well as temperature (Cavosie et al. 2000). Quartz + kyanite and quartz + sillimanite pairs from British Columbia yield precise temperatures from quartz nodules, but lower reset temperatures from assemblages in poly-mineralic pelites (Table 4). Fast Grain Boundary diffusion calculations based on the difference between RAM and pelite temperatures suggest that water activity was low during slow retrograde cooling (Ghent and Valley 1998). Other applications of quartz-aluminosilicate fractionations include migmatitic, amphibolite or granulite facies gneisses (van Haren et al. 1996, Kohn et al. 1997, Vannay et al. 1999, Moecher and Sharp 1999), and eclogites (Sharp et al. 1992, 1993; Rumble and Yui 1998, Zheng et al. 1998, 1999). Moecher and Sharp (1999) compared the results of aluminosilicate-quartz and aluminosilicate-garnet thermometry in pelites and report variable retrograde resetting as predicted by diffusion modeling. Magnetite-quartz

The magnetite-quartz pair is the most commonly applied oxygen isotope thermometer. It is very promising as a RAM thermometer below TC(Mt) or when used in low to moderate grade metamorphic, rapidly cooled, or very dry rocks. It is frequently discussed in reviews of isotope thermometry (O’Neil and Clayton 1964, Chiba et al. 1989, Gregory et al. 1989, Zheng and Simon 1991). The fractionation is large yielding good temperature sensitivity and the effects of solid solution and crystal chemistry are relatively small (Ti, Fe3+ in magnetite, Bottinga and Javoy 1975, Zheng and Simon 1991; SiO2 polymorphs, Kawabe 1978, Zheng 1993b; spinel structure, Zheng 1995). Rumble (1978) measured Mt-Qt fractionations from eight amphibolite facies quartzites on the summit of Black Mountain, New Hampshire. These closely-spaced samples were subjected to the same P-T-time conditions. All but one sample contains at least 80% quartz and smaller amounts of magnetite. Coexisting minerals indicate a peak metamorphic temperature of 530°C: kyanite + staurolite + chloritoid + chlorite + muscovite + quartz + magnetite ± ilmenite ± garnet. The fractionations are self-consistent (9.46±0.26‰) yielding nearly parallel tie lines (Fig. 14). The oxygen isotope temperatures range from 530-561°C (542±11, Table 5). This study is an outcrop test and demonstrates the accuracy and precision obtainable from Mt-Qt when applied to appropriate rocks for a RAM thermometer. The magnetite/hematite-quartz thermometer has been extensively applied to banded

Stable Isotope Thermometry at High Temperatures

389

Figure 14. Plot of δ18O(magnetite) vs. δ18O(quartz) for amphibolite facies quartzites from Black Mtn., New Hampshire. The nearly parallel tie-lines indicate that fractionations are self-consistent for these eight samples from one outcrop. The average oxygen isotope temperature of 542°C is in excellent agreement with that from petrology (530oC). (from Rumble (1978).

Table 5. Comparison of Δ18O(Qt-Mt) temperature estimates from quartzite, granitic gneiss, and banded iron formation. Amphibolite facies Black Mtn. quartzite yields accurate and precise RAM temperatures. Granulite facies temperatures are often low relative to petrologic estimates. Reference, Location

1. Shuksan, N. Cascades 2. Black Mtn., N.H., Qtzt, 3. Isua, SW Greenland 4. Minas Gerais, Brazil

Average Δ(Qt-Mt)

13.8±1.50 9.46±0.26 11.62±0.42 10.79±0.39 8.14±0.77 6.32±0.41 8.89±2.47 5. Ruby Range, Mont., Kelly 8.42±0.19 Carter Creek 9.67±0.43 6. Adirondack Mts, unsheared 6.93±0.58 sheared 8.8±1.22 7. Wind River Range, Wyoming 9.24±0.93

a

Ave T n

°C

6 8 12 16 38 9 16 3 3 11 11 16

402±37 542±11 463±13 491±14 606±40 724±33 568±88 591±10 533±18 680±38 603±71 552±44

Comments BSF, lawsonite, pumpellyite AF, 530oC, 500 × 100 m, RAM AF, BIF, 1 × 2 km GSF, BIF, s1 onlyb AF, BIF, s1 onlyb GF, BIF, s1 onlyb GF, BIF, s1+s2+s3c GF, 745±50oC, BIF, 1 km2 GF, 675±45oC, BIF, 9 km2 GF, 675-700oC, GG GF, 675-700oC, GG GF, 750±50oC, BIF

References: 1. Brown and O’Neil (1982), 2. Rumble (1978); 3. Perry et al. (1978); 4. Muller et al. (1986b); 5. Dahl (1979); 6. Cartwright et al. (1993); 7. Sharp et al. (1988); Sharp and Essene (1991). Footnotes: a calculated using the calibrations of Chiba et al. (1989); b data only for samples of magnetite/hematite showing primary schistosity s1; c data for samples of magnetite/hematite with variable schistosity s1, s2, or s3 Abbreviations: AF = amphibolite facies; BIF = banded iron formation; BSF = blueschist facies; GF = granulite facies; GG = granitic gneiss or charnockite; GSF = greenschist facies; RAM = refractory accessory mineral thermometer; Qtzt = quartzite.

iron formations in order to determine conditions of deposition, diagenesis and metamorphism: Animikie Basin, U.S. (Clayton and Epstein 1958, James and Clayton 1962); Duluth gabbro contact aureole, U.S. (James and Clayton 1962, Perry and Bonnichsen 1966); Hamersley Basin, Western Australia (Becker and Clayton 1976); Isua, SW Greenland (Perry et al. 1978); Krigoy Rog, Ukraine (Perry and Ahmad 1981); Mesabi

390

Valley

Range, U.S. (James and Clayton 1962, Perry et al. 1973); Minas Gerais, Brazil (Hoefs et al. 1982, Muller et al. 1986a,b); Ruby Range, Montana (Dahl (1979); Urucum area, Brazil (Hoefs et al. 1987); and the Wind River Range, Wyoming (Sharp et al. 1988, Sharp and Essene 1991). These data include samples from different metamorphic grades: unmetamorphosed to sub-greenschist (Animikie, Hamersley, Mesabi, Urucum); greenschist facies (Krivoy Rog, Minas Gerais); amphibolite facies (Isua, Minas Gerais); granulite facies (Minas Gerais, Ruby Range, Wind River); and contact metamorphism (Duluth gabbro aureole). Reexamination of these samples, using more recent criteria for thermometry, can be expected to yield improved accuracy. The temperatures found by these studies should be recalculated with newer experimental data (Chiba et al. 1989, Clayton et al. 1989). Furthermore, it is possible that isotope heterogeneity exists at the scale of the large samples analyzed, and that equilibrated domains can be identified and analyzed with microanalysis. Magnetite-quartz thermometry has also been applied in igneous rocks (Hildreth et al. 1984, Bindeman and Valley 2001) and high grade gneisses (Fourcade and Javoy 1973, Shieh and Schwarcz 1974, Shieh et al. 1976, Li et al. 1991, Cartwright et al. 1993, Hoffbauer et al. 1994, Farquhar et al. 1996), blueschists (Brown and O’Neil 1982), and greenschist or amphibolite facies pelites (Schwarcz et al. 1970, Hoernes and Friedrichsen 1974, Kerrich et al. 1977, Goldman and Albee 1977). There are several cautionary notes for thermometry involving Fe-Ti oxides and quartz. Minerals may differentially exchange with circulating fluids as clearly demonstrated by Gregory et al. (1989) with δ18O(Qt) vs. δ18O(Mt) plots. For Hamersley and Mesabi, it is proposed that sedimentary Fe-oxides exchanged with fluids during recrystallization to magnetite, while quartz remained unaffected. The calibration of 1000lnα18O(Qt-Mt) vs. T is uncertain at low temperatures. Values of Δ(Mt-hematite) and Δ(ilmenite-Mt) are small, but not insignificant (Bottinga and Javoy 1975, Zheng 1991, Zheng and Simon 1991). In Ti- or Fe+3-rich oxides, exsolution effects are minimized for lower grade rocks, but may be significant in granulites (Bohlen and Essene 1977, Farquhar and Chacko 1994). Fine grain size has sometimes precluded complete mineral separation and various projection schemes have been applied to banded iron formation (see, Yapp 1990). The moderate diffusion rates of oxygen in both quartz and magnetite, and measured fractionations indicate that temperatures will be reset above TC, for granulite and upper amphibolite facies rocks (Chiba et al. 1989, Valley and Graham 1991, Sharp 1991, Sharp and Essene 1991, Eiler et al. 1993, 1995a,b). In addition to diffusive processes that operate in all samples, both magnetite and quartz can exchange δ18O by recrystallization or crack-healing (Valley and Graham 1993, 1996; Eiler et al. 1995a, Sitzman et al. 2000). Likewise, different generations of quartz are recognized from textures in Archean Onvervacht cherts and small differences in δ18O have been measured (Knauth and Lowe 1978). In spite of the many potential pitfalls, Mt-Qt is an accurate and reliable thermometer when carefully applied to appropriate rocks in an appropriate temperature regime. Samples should be selected with regard to diffusive exchange, discussed above. Temperature estimates are consistent with petrology for blueschist, greenschist and lower amphibolite facies samples in Table 5. In spite of the moderate TC for both quartz and magnetite (Table 3), some granulite facies gneisses preserve peak metamorphic temperatures (Muller et al. 1986b, Cartwright et al. 1993) and others yield temperatures that are reset, but higher than TC (Sharp et al. 1988). However, most granulite and upper amphibolite facies samples yield temperatures that are too low in comparison to petrologic thermometers, though reset results can be precise (Table 5). The variable retention of peak temperature is predicted if water fugacity is low (promoting slow

Stable Isotope Thermometry at High Temperatures

391

diffusion) in some rocks during cooling, but high in others (Sharp et al. 1988, Cartwright et al. 1993, Edwards and Valley 1998). Improved results have been obtained through careful attention to rock and mineral fabric in regionally metamorphosed terranes (Muller et al. 1986a, Sharp et al. 1988, Cartwright et al. 1993). Crystal dislocations in magnetite that can facilitate retrograde oxygen exchange are easily seen in polished thin sections that have been etched in HCl (Valley and Graham 1993, Eiler et al. 1995b, Sitzman et al. 2000). Healed microfractures in quartz, if present, are often imaged by cathodoluminescence using a SEM or electron microprobe, though sensitivity varies greatly with instrument and operating conditions (Valley and Graham 1996). Rutile-quartz

The Ru-Qt thermometer has the same theoretical advantages as Mt-Qt. The temperature coefficient of fractionation is large and solid-solution is minor. The greatest limitation of this system may be the rate of oxygen diffusion (Table 3). Quartz-rutile pairs have been measured most commonly from eclogites (Vogel and Garlick 1970, Desmons and O’Neil 1978, Matthews et al. 1979, Agrinier et al. 1985, Agrinier 1991, Sharp 1995, Rumble and Yui 1998, Zheng et al. 1998, 1999). A few analyses are reported for blueschists and pelites (Matthews and Schliestedt 1984, Sharp 1995), and nelsonite (Addy and Garlick 1974). Vogel and Garlick (1970) report very high precision: Δ(Qt-Ru)= 6.46±0.05‰ for 5 unrelated type B eclogites (609±4°C, Matthews 1994) and 7.30‰ for one type C (556°C). However, other studies are less sucessful. Sharp (1995) reports analyses from five granulite facies samples from terranes where the reported temperatures average 755°C and Δ(Qt-Ru) yields average temperatures of 567±150°C. The only granulite that yields above 700°C is a rapidly quenched xenolith from La Joya Honda maar. Plots of Δ(Qt-Ru) vs. Δ(Qt-garnet) and Δ(Qt-Ru) vs. Δ(Qtkyanite) for the data referenced above show considerably larger scatter than Δ(Qt-garnet) vs. Δ(Qt-kyanite) for the same rocks, suggesting that in spite of the smaller temperature sensitivity, these other pairs may be more reliable. The effect of diffusion cannot be predicted for most of these samples without information on modes and grain size that is not published. Furthermore, rutile has the unusual property that hydrous experiments yield slower diffusion coefficients than dry experiments in the same lab (Moore et al. 1998). Since quartz is a minor phase in many of these mafic rocks, it is likely that variable resetting to lower temperatures has occurred by exchange with micas or chlorite. In a few samples with high Qt-Ru temperatures, it is possible that retrograde fluids were present. Rutile is common in certain quartz veins, quartzites, and some massif-type anorthosites. These assemblages should be sought as a test of the RAM thermometers. Within the temperature range dictated by diffusion in rutile (Table 3) it is predicted to be highly accurate and precise. CARBON ISOTOPE THERMOMETRY Calcite-graphite

Cc-Gr is the most commonly applied RAM thermometer (Table 6). The percentage of graphite is generally less than 1% in marble, other carbon-bearing phases such as scapolite or dolomite are usually minor in abundance, and the diffusion rate of carbon is very slow. Furthermore, Δ(Cc-Scap) and Δ(Cc-Dol) are small at high temperatures, minimizing the effect of neglecting small amounts of these minerals. With routine care, this is an easily applied and accurate thermometer for high-grade marbles. A number of empirical, experimental or theoretical calibrations of Δ13C(Cc-Gr) vs. T

392

Valley

Table 6. Comparison of Δ13C(Cc–Gr) temperature estimates (oC) from granulite and amphibolite facies terranes. Average Ref. Location Δ(Cc–Gr) 1. Anabar Shield, Russia 3.83 2. Bohemian Massif 4.81 3. Central Adirondack Mtns. 3.44a near anorthosite 3.02 4. Central Adirondack Mtns 3.49 5. NW Adirondacks 4.08 6. Cucamonga terrane, California 3.54 7. Franklin marble, amph. facies 3.60 gran. facies 3.09 8. Gour Oumelalen, Algeria 3.45 9. Hida Belt, Japan 3.46 10. Ivrea Zone, Italy 4.9-2.3 11. Kamioka area, Japan 3.54 12. Kerala Belt, S. India 3.33 13. Kurobegawa area, Japan 5.77 14. Lutzow Holm Bay, E Antarctica 2.85 15. Madurai, southern India 2.67 16. Panamint Mtns, marbles 6.40 schists 8.78 17. Sanbagawa terrain, Japan 8.37 18. Skallen marble, E. Antarctica 2.85 19. Southern Grenville Province, ON 6.88 20. Sri Lanka, SW Group marbles 4.49 21. Sri Lanka, granulite facies 3.79 altered calcite 2.93 22. Tudor Gabbro aureole, 1km 7.32 23. Madurai, Tamil Nadu 4.06

n 4 9 38 17 10 89 10 3 3 10 5 14 40 10 9 2 18 17 18 15 5 31 14 19 6 17 8 3

Cc-Gr Ave T’s b K+V D+V 691 677 590 613 744 707 813 743 737 703 661 659 746 705 722 695 801 737 743 706 742 706 580-970 600-800 729 699 760 716 513 560 845 759 881 776 473 530 364 441 379 455 844 758 447 510 618 632 696 680 829 751 636 644 424 492 662 660

Independent estimate T oCc 500°C), crystallinity increases at the atomic-scale from moderately well-organized rims on grains with immature cores to well-organized crystalline flake graphite (Buseck and Huang 1985). Graphitization in marble is also accompanied by carbon isotope exchange. Carbon diffusion in graphite is very slow (Thrower and Mayer 1978) and TC » 800°C (Table 3). High TC is further indicated by isotopically zoned flakes of graphite that were not homogenized by granulite facies metamorphism (Kitchen and Valley 1995, Farquhar et al. 1999, Satish-Kumar et al. 2000). Thus, carbon isotope exchange between graphite and carbonate is sluggish, rate-limited by graphite, and can only occur upon crystallization or recrystallization of graphite. In contrast, carbon diffusion in calcite is relatively rapid. Accordingly, Cc-Gr thermometry should only be attempted with visible flakes of graphite and the presence of growth-zoning within flakes should be evaluated. Figure 16. Measured values of Δ(Cc-Gr) versus metamorphic temperature independently determined by petrologic thermometers. Data are from: Valley and O’Neil (1981), open triangles; Wada and Suzuki (1983), asterisks; Kreulen and van Beek (1983), squares; Morikiyo (1984), circles; and Dunn and Valley (1992), solid triangles and heavy curved line. The unmetamorphosed calcite-organic matter fractionation is from Eichmann and Schidlowski (1975). Samples show a successive approach to equilibrium at higher temperatures rate-limited by slow exchange in graphite (from Dunn and Valley 1992).

A number of studies have measured Cc-Gr fractionations that appear too large in relation to equation (10), resulting in temperatures that are lower than expected (Hoefs and Frey 1976, Kreulen and van Beek 1983, Wada and Suzuki 1983, Morikiyo 1984, Wada et al. 1984, Schrauder et al. 1993, Bergfield et al. 1996). These studies support the hypothesis that either the calibration has a steeper slope at low temperatures than is shown in Figure 15 or that the samples are not fully equilibrated. The data in and 17 show that exchange begins at 300°C, but that below 500-600°C, large values of Δ13C(Cc-Gr) are seen that are highly variable indicating incomplete exchange. Equilibration is not assured below 650°C. Dunn and Valley (1992) found that samples with flakes of graphite (vs. less mature carbonaceous material) plot along the bottom of this envelope of fractionations (Fig. 16) and they suggested that the lowest values were equilibrated. Accordingly, the heavy line in Figure 16 is their calibration (Eqn. 11). For these reasons, it is unlikely that values of Δ13C(Cc-Gr) > 8‰ represent equilibrium and values between 5 and 8‰ should be interpreted with extra caution. Once coarsening and recrystallization stops, graphite flakes will preserve isotope composition due to slow diffusion. Intracrystalline zoning of δ13C has been detected in some crystals by most studies that have evaluated it. This has been demonstrated by delamination of single flakes at the 20-100 μm-scale (Wada 1988, Arita and Wada 1990, Kitchen and Valley 1995, Satish-Kumar 2000), by ion microprobe (Farquar et al. 1999) or, more simply, by analysis of large vs. small flakes from the same rock (Kitchen and Valley 1995).

Stable Isotope Thermometry at High Temperatures

395

Figure 17. Carbon isotope fractionations as a function of metamorphic grade. (17A) Compilation of measured values of Δ(CcGr) shows scatter and disequilibrium at greenschist facies and lower temperature conditions. Many amphibolite facies and most granulite facies samples show a tight clustering of values consistent with isotope equilibration above 600oC. Values in black are from the Adirondacks (from Kitchen and Valley 1995). (17B) Values of δ13C for a Liassic black shale formation (Hoefs and Frey 1976) showing successive approach to equilibrium at maximum T = 500-600°C (from Sharp et al. 1995).

396

Valley

Graphite inclusions wholly within silicate minerals have also been found to have slightly lower δ13C than graphite in the calcite matrix suggesting that they were armored against prograde exchange (Wada and Suzuki 1983). This effect will not affect thermometry if care is taken to only select graphite that is disaggregated by acid dissolution of carbonate. Conversely, the analysis of selected inclusions from different minerals, or from different positions within a zoned mineral, may provide temperature history of prograde mineral growth. Skeletal graphite or flakes with etched or pitted surfaces are frequently seen, suggesting retrograde recrystallization. Skeletal graphites with larger Δ13C(Cc-Gr) than metallic flakes support this interpretation (Weis 1980, Arita and Wada 1990). Van der Pluijm and Carlson (1989) determined temperatures were less than 450°C during mylonitization along the Bancroft shear zone in the Grenville Province by reequilibration of Δ13C(Cc-Gr) in 650-700°C marbles. Likewise, graphite spheres with radial texture have been described in nearby marble and Cc-Gr thermometry suggests lower temperature growth (Jaszczak and Robinson 1998; S. Dunn, unpublished 1999). Samples of charnockite from southern India have been shown to contain three generations of graphite varying by up to 10‰ in δ13C by ion microprobe (Farquhar et al. 1999). Carbonate/graphite ratio

Accurate thermometry is most likely if the ratio of carbon-in-carbonate to carbon-ingraphite is high. The mode effects on diffusive resetting, described above, are a criterion for RAM thermometry. Also, retrograde precipitation of carbonate is common and may be more significant in carbonate-poor lithologies. Considering the concentration of carbon in each mineral, 100% in graphite vs. 12 wt % in calcite, mode ratios on the order of 100/1 for carbonate/graphite are prudent. This is seldom a concern in marbles where 0.1 to 1.0 vol % graphite is typical, but care should be considered for graphite-rich or carbonate-poor lithologies. Elsenheimer (1988, 1992) analyzed 11 Cc-Gr pairs from calcite-poor calcsilicates ( -10‰, this value results from metamorphic exchange with calcite. In all rocks, the protolith carbonaceous matter is estimated at δ13C < -25, indicative of a biogenic origin. Likewise, Morrison and Barth (1993) and Bergfield et al. (1996) show that mass-balance and exchange with carbonate controls δ13C(Gr) in marbles. Adirondack Mountains—A case study

Graphite-bearing calcite marbles are common throughout the upper amphibolite facies NW Adirondack Lowlands and locally in the granulite facies central Adirondack Highlands. Figure 19A shows the results of Cc-Gr thermometry for 89 samples from the NW and 55 from the central Adirondacks (Kitchen and Valley (1995). All graphites were flakes freed by dissolution of carbonate. Five outcrop tests were made and reproducibility of temperatures is ±10-20°C except where calcites are heterogeneous (Table 2). The 55 granulite facies samples include 17 that are from near the plutonic contacts of massif-type anorthosite bodies. These samples contain zoned graphite crystals with higher δ13C cores that record the pre-regional metamorphism igneous contact temperatures. The data for the NW Adirondacks show that the central zone of the region (B in Fig. 19B) experienced systematically lower temperatures, averaging 640°C and that temperatures increase systematically towards transitions to granulite facies conditions to the SE across the Carthage-Colton Line and to the NW across the St. Lawrence River. A temperature of 675°C is recorded for the orthopyroxene isograd to the SE. The two 675°C isotherms parallel the strike of dominant lithologies in the area and are normal to the NW vergence, in good agreement with recent petrologic thermometry (Liogys and

Stable Isotope Thermometry at High Temperatures

399

Jenki Figure 19. Carbon isotope thermometry for Adirondack marbles. (19A) Plot of δ13C(calcite) vs. δ13C(graphite). Open symbols are for upper the amphibolite facies NW Adirondacks and closed symbols are for the granulite facies Adirondack Highlands. (19B) Values of Δ(Cc-Gr) in per mil and resulting isopleths for 675oC parallel the strike of the terrane and, show lower metamorphic temperatures centered on Gouverneur and increasing towards granulite facies rocks to the SE and to NW. Note outcrop tests at Fish Creek (FC), the Train Wreck (TW) and the Valentine Mine (VM) (from Kitchen and Valley 1995).

Jenkins 2000). No zones of higher temperature were found at exposed igneous contacts, though the locally low values of Δ13C(Cc-Gr) 5 km W of Canton may be an unrecognized contact aureole. The Cc-Gr RAM thermometer provides the best estimates for metamorphic temperature in this terrane. SULFUR ISOTOPE THERMOMETRY

Several common sulfide mineral pairs have fractionations that are calibrated and appropriate for thermometry, especially in sulfide deposits. However thermometry has often been disappointing, and sulfur isotope studies have tended to concentrate on elucidating fluid conditions, sources of sulfur, and ore mineral paragenesis (Ohmoto 1986). Recent microanalysis by laser probe and ion microprobe documents one of the main problems. Sulfides forming at low to intermediate temperatures show extreme zonation of δ34S, over 50‰ in less than 100 μm in some diagenetic pyrites (McKibben and Riciputi 1998) and 4-5‰/cm in a hydrothermal black smoker vent (Shanks et al. 1998).

400

Valley

Crowe (1994) analyzed coexisting pyrite-chalcopyrite pairs from the stockwork feeder zones of several volcanogenic massive sulfide deposits. Individual 200-μm diameter grains were analyzed by laser. Sulfides that were in grain to grain contact were found to have exchanged during greenschist to upper amphibolite facies metamorphism and during cooling, while sulfides that were isolated by quartz matrix preserved original hydrothermal δ34S values. Temperatures of 170-424°C were calculated for three different hydrothermal systems. Thus, reliable thermometry will require assessment of sub-mm scale isotope zonation and exchange kinetics in sulfide minerals. SKARNS

Skarns and hydrothermal veins associated with shallow igneous activity offer excellent opportunities for stable isotope thermometry because minerals commonly coprecipitate directly from a fluid and cool quickly. Minerals that are not zoned in chemical composition are most promising. Tests can be applied for mineral zoning or precipitation from evolving fluids. Bowman (1998) reviewed stable isotope studies of skarns (Table 7). In a number of skarns, temperatures are reported that are generally consistent with the results of phase equilibria and fluid inclusion thermometry. However, minerals that have not coprecipitated do not typically yield self-consistent temperature estimates with the different techniques. Stable isotope concordance is also a good test. For instance, quartz, calcite and magnetite appear texturally to be co-precipitated and were analyzed from ten rocks in the Hanover Zn-Pb skarn (Table 7, #8, Turner and Bowman 1993). The Qt-Mt temperatures are in good agreement with independent thermometry, but the Cc-Qt and Cc-Mt temperatures are too low and too high, respectively (Table 7) because of late-stage depletion of 18O/16O in calcite (Bowman 1998). Stable isotope zonation at the cm-scale is well documented in skarns, but few samples have been tested at mm to μm-scale (Jamtveit and Hervig 1994, Bezenek et al. 1995, Clechenko and Valley 2000). A careful study would include petrography and imaging, and other tests of temperature reliability. Microanalysis should be employed when other tests fail. It offers a powerful new tool for resolving the thermal and fluid evolution of skarns. ONE-MINERAL THERMOMETERS

Isotope fractionation can occur among crystallographically distinct sites within single minerals creating a potential one mineral thermometer. However, techniques for extraction and analysis of the isotope ratio of a specific site vary greatly in difficulty and reliability, and more experimental calibrations of fractionation are needed. Specific studies for oxygen isotopes include: clays (Hamza and Epstein 1980, Bechtel and Hoernes 1990, Sheppard and Gilg 1996); phyllosilicates (Savin and Lee 1988); analcime NOTES for Table 7 (next page) 1

Mineral abbreviations: Am = amphibole, Anh = anhydrite, Bt = biotite, Cal = calcite, Ccp = chalcopyrite, Chl = chlorite, Ep = epidote, Grt = garnet, Gn =galena, Hem = hematite, Mag = magnetite, Ms = muscovite, Pl = plagioclase, Px = pyroxene,. Py = pyrite, Qtz = quartz, Sp = sphalerite. 2 Number of pairs analyzed in parentheses. 3 Pressure (MPa) used in study for calculating phase equlibria and for correcting fluid-inclusion microthermometry data.. *References: 1. Taylor & O’Neil 1977; 2. Shelton & Rye 1982; Shelton 1983; 3. Bowman et al. 1985a; 4. Jamtveit et al. 1992a,b; 5. Bowman et al. 1985b; 6. Brown et al. 1985; 7. Layne & Spooner 1991; Layne et al. 1991; 8. Turner & Bowman 1993; 9. Kemp 1985; 10. Shimizu & Iiyama 1982.

10.

9.

8.

7.

6.

5.

4.

3.

2.

Nakatatsu, Japan

Alta, Utah

Hanover, New Mexico

JC, Yukon

Pine Creek, California

Elkhorn, Montana

Oslo Rift, Norway

Cantung, NWT

Mines Gaspe, Quebec

Osgood Mtns., Nevada

Zn, Pb

Cu

Zn, Pb

Sn

W

-

W, Zn Cu

W

Cu

W Cu, Au

II

I II

III

I

III IV V

I

I

II

I II

II III

I II

Stage

329

538

391 432 625 358 172 401

460 446 441

494

497

360

478 505

412 688

381 523 636

364

595 Reversal

405 506 1128 483 377 533

504 446 441

560

660

360

500 525

552 801

640 540 1000

278

500

377 337 353 291 -3 336

416 446 441

400

420

360

460 485

366 550

330 480 505

Isotope temperature Ave. Max. Min.

Sp-Gn (4)

Px-Mag (5) Qtz-Cal (5)

Grt-Mag (2) Cal-Ep,Chl (13) Qtz-Cal,Ep (9) Qtz-Mag (12) Qtz-Cal (10) Cal-Mag (10)

Qtz-Ep (2) Qtz-Bt (1) Cal-Ms (1)

Qtz-Px (8)

Pl-Px (11)

Qtz-Grt (1)

Qtz-Px (6) Qtz-Bt (2)

Py-Ccp (6) Anh-Py, Ccp (20)

Qtz-Grt (11) Qtz-Am (6) Qtz-Cal (7)

Skarn *(reference)

390

575 395

315

380 0

350

350

>580 445

375

400

500 560

400

410 400

425

380 345

255

305

410 430

300

360 350

200

Fluid inclusion Ave. Max. Min.

355

490

340

Phase equil.

425

550 450 450

590

560

380

575 475

500

575 475

Type

Table 7. Comparison of temperature estimates (°C) from stable isotopes, phase equilbria, and fluid inclusions in skarns (from Bowman 1998).

30

150 150

40

40

75 75

150

100

50

100 100

30

150 150

PCorr (Mpa)

Stable Isotope Thermometry at High Temperatures 401

402

Valley

(Karlsson and Clayton 1990, Cheng et al. 2000); micas and chlorite (Hamza and Epstein 1980); scapolite (Moecher et al. 1984); alunite (Ustinov and Grinenko 1985, O’Neil and Pickthorn 1988, Rye et al. 1992); sahaite and spurrite (Ustinov and Grinenko 1985); and apatites (Shemesh et al. 1983, 1988, Ustinov and Grinenko 1985, Zheng 1996). Such studies hold great promise, especially in sedimentary rocks where co-precipitated minerals are rare. ACKNOWLEDGMENTS

I thank James R. O’Neil for introducing me to the rigor and the fun of stable isotope geochemistry, and for many years of friendship and inspiration. Much of the research reported here was supported by NSF and DOE. Figures were redrafted by Mary Diman. John Eiler, James Farquhar, Yaron Katzir, William Peck, and Zach Sharp are thanked for helpful reviews. REFERENCES Addy SK, Garlick GD (1974) Oxygen isotope fractionation between rutile and water. Contrib Mineral Petrol 45:119-121 Agrinier P (1991) The natural calibration of 18O/16O geothermometers: Application to the quartz-rutile mineral pair. Chem Geol 91:49-64 Agrinier P, Javoy M, Smith DC, Pineau F (1985) Carbon and oxygen isotopes in eclogites, amphibolites, veins and marbles from Western Gneiss Region, Norway. Isotope Geosci 52:146-162 Anderson AT (1967) The dimensions of oxygen isotope equilibrium attainment during prograde metamorphism. J Geol 75:323-332 Arita Y, Wada H (1990) Stable isotopic evidence for migration of metamorphic fluids along grain boundaries of marbles. Geochem J 24:173-186 Baker AJ (1988) Stable isotope evidence for limited fluid infiltration of deep crustal rocks from the Ivrea Zone, Italy. Geology 16:492-495 Bechtel Z, Hoernes S (1990) Oxygen isotope fractionation between oxygen of different sites in illite minerals: A potential single-mineral thermometer. Contrib Mineral Petrol 104:463-470 Becker RH, Clayton RN (1976) Oxygen isotope study of a Precambrian banded iron-formation, Hamersley Range, Western Australia. Geochim Cosmochim Acta 40:1153-1165 Bergfeld D, Nabelek PI, Labotka TC (1996) Carbon isotope exchange during polymetamorphism in the Panamint Mountains, California, USA. J Metamor Geol 14:199-212 Bestmann M, Kunze K, Matthews A (2000) The evolution of calcite marble shear zone complex on Thassos island, Greece; Microstructural and textural fabrics and their kinematic significance. J Struc Geol 22:1789-1807 Bezenek SR, Crowe DE, Riciputi LR (1995) Evidence of protracted growth history of skarn garnet using SIMS oxygen isotope, trace element and rare earth element data. Geol Soc Am Abstr Prog 27:67 Bigeleisen J, Mayer MG (1947) Calculation of equilibrium constants for isotopic exchange reactions. J Chem Phys 15:261-267 Bigeleisen J, Perlman ML, Prosser HC (1952) Conversion of hydrogenic materials for isotopic analysis. Analyt Chem 24:1356 Bindeman IN, Valley JW (2000) Formation of low-δ18O rhyolites after caldera collapse at Yellowstone, Wyoming, USA. Geology 28:719-722 Bindeman IN, Valley JW (2001) Oxygen isotope study of phenocrysts in zoned tuffs and lavas from Timber Mountain/Oasis Valley caldera complex: Generation of large volumes of low-δ18O. Contrib Mineral Petrol (in review) Bohlen SR, Essene EJ (1977) Feldspar and oxide thermometry of granulites in the Adirondack Highlands. Contrib Mineral Petrol 62:153-169 Bottinga Y, Javoy M (1973) Comments on oxygen isotope geothermometry. Earth Planet Sci Letters 20:250-265 Bottinga Y, Javoy M (1975) Oxygen Isotope Partitioning Among the Minerals in Igneous and Metamorphic Rocks. Rev. Geophys. Space Phys 13:401-418 Bottinga Y, Javoy M (1987) Comments on stable isotope geothermometry: The system quartz-water. Earth Planet Sci Letters 84:406-414 Bowman JR (1998) Stable-Isotope Systematics of Skarns. In Lentz DR (ed) Mineral Assoc Canada Short Course Series: Mineralized Intrusion-Related Skarn Systems, p 99-145

Stable Isotope Thermometry at High Temperatures

403

Bowman JR, Covert JJ, Clark AH, Mathieson GA (1985a) The Can Tung scheelite skarn orebody, Tungsten, Northwest Territories, Canada: Oxygen, hydrogen and carbon isotope studies. Econ Geol 80:1872-1895 Bowman JR, O'Neil JR, Essene EJ (1985b) Contact skarn formation at Elkhorn Montana. II: Origin and evolution of C-O-H skarn fluids. Am J Sci 285:621-660 Brady JB, Cheney JT, Larson Rhodes A, Vasquez A, Green C, Duvall M, Kogut A, Kaufman L, Kovaric D (1998) Isotope geochemistry of Proterozoic talc occurrences in Archean marbles of the Ruby Mountains, southwest Montana, U.S.A. Geol Materials Res 1:1 [ http://www.minsocam.org ] Brown EH, O'Neil JR (1982) Oxygen Isotope Geothermometry and Stability of Lawsonite and Pumpellyite in the Shuksan Suite, North Cascades, Washington. Contrib Mineral Petrol 80:240-244 Brown PE, Bowman JR, Kelly WC (1985) Petrologic and stable isotope constraints on the source and evaluation of skarn-forming fluids at Pine Creek, California. Econ Geol 80:72-95 Buseck PR, Huang BJ (1985) Conversion of carbonaceous material to graphite during metamorphism. Geochim Cosmochim Acta 49:2003-2016 Cartwright I, Valley JW, Hazelwood A (1993) Resetting of oxybarometers and oxygen isotope ratios in granulite facies orthogneisses during cooling and shearing, Adirondack Mountains, New York. Contrib Mineral Petrol 113:208-225 Cavosie AJ, Sharp ZD, Selverstone J (2000) Application of a stable isotope geobarometer: Co-existing aluminosilicates in isotopic equilibrium from the Northern Colorado front range. Geol Soc Am Annual Meeting, p A-115 Cawley JD (1984) Oxygen Diffusion in Alpha Alumina. PhD Thesis, Case Western Reserve University, Cleveland, Ohio Chacko T, Mayeda TK, Clayton RN, Goldsmith JR (1991) Oxygen and carbon isotope fractionations between CO2 and calcite. Geochim Cosmochim Acta 55:2867-2882 Chamberlain CP, Conrad ME (1991) The relative permeabilities of quartzites and schists during active metamorphism at mid-crustal levels. Geophys Res Letters 18:959-962 Chamberlain CP, Conrad ME (1992) Oxygen-isotope zoning in garnet: A record of volatile transport. Geochim Cosmochim Acta 57:2613-1619 Chamberlain CP, Ferry JM, Rumble D (1990) The effect of net-transfer reactions on the isotopic composition of minerals. Contrib Mineral Petrol 105:322-336 Cheng X, Zhao P, Stebbins JF (2000) Solid state NMR study of oxygen site exchange and Al-O-Al site concentration in analcime. Am Mineral 85:1030-1037 Chiba H, Chacko T, Clayton RN, Goldsmith JR (1989) Oxygen isotope fractionations involving diopside, forsterite, magnetite, and calcite: Application to geothermometry. Geochim Cosmochim Acta 53:29852995 Clayton RN (1981) Isotopic thermometry. In Newton RC, Navrotsky A, Wood BJ (ed) Advances in Physical Geochemistry. Springer-Verlag, Berlin, p 85-109 Clayton RN, Epstein S (1958) The relationship between O18/O16 ratios in coexisting quartz, carbonate, and iron oxides from various geological deposits. J Geol 66:352-373 Clayton RN, Mayeda TK (1963) The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotopic analysis. Geochim Cosmochim Acta 27:43-52 Clayton RN, Kieffer SW (1991) Oxygen isotopic thermometer calibrations. In Taylor HP, O'Neil JR, Kaplan IR (ed) Stable Isotope Geochemistry: A Tribute to Samuel Epstein. Geochem Soc Spec Publ 3:3-10 Clayton RN, Goldsmith JR, Mayeda TK (1989) Oxygen isotope fractionation in quartz, albite, anorthite and calcite. Geochim Cosmochim Acta 53:725-733 Clayton RN, Goldsmith J, Karel KJ, Mayeda T, Newton RC (1975) Limits on the effect of pressure on isotopic fractionation. Geochim Cosmochim Acta 39:1197-1201 Clechenko CC, Valley JW (2000) Oscillatory zoned skarn garnet adjacent to massif anorthosite, Willsboro wollastonite mine, N.E. Adirondack Mts, N.Y. Geol Soc Am Abstr Prog, p A-295 Coghlan RAN (1990) Studies in diffusional transport: Grain boundary transport of O in feldspars, diffusion of O, strontium, and the REEs in garnet and thermal histories of granitic intrusions in south-central Maine using O isotopes. PhD Thesis, Brown University, Providence, Rhode Island Craig H (1953) The geochemistry of the stable carbon isotopes. Geochim Cosmochim Acta 3:53-92 Crank J (1975) The Mathematics of Diffusion. Clarendon Press, Oxford, UK Crawford WA, Valley JW (1990) Origin of graphite in the Pickering gneiss and the Franklin marble, Honey Brook Upland, Pennsylvania Piedmont. Geol Soc Am Bull 102:807-811 Criss RE (1999) Principles of Stable Isotope Distribution. Oxford Press, Oxford, UK Criss RE, Gregory RT, Taylor HP (1987) Kinetic theory of oxygen isotopic exchange between minerals and water. Geochim Cosmochim Acta 51:1099-1108

404

Valley

Crowe DE (1994) Preservation of original hydrothermal δ34S values in greenschist to upper amphibolite volcanogenic massive sulfide deposits. Geology 22:873-876 Crowe DE, Valley JW, Baker KL (1990) Microanalysis of sulfur isotope zonation by laser microprobe. Geochim Cosmochim Acta 54:2075-2092 Dahl PS (1979) Comparative geothermometry based on major-element and oxygen isotope distributions in Precambrian metamorphic rocks from southwestern Montana. Am Mineral 64:1280-1293 Deines P (1977) On the oxygen isotope distribution among mineral triplets in igneous and metamorphic rocks. Geochim Cosmochim Acta 41:1709-1730 Deloule E, Allegre CJ, Doe B (1986) Lead and sulfur isotope microstratigraphy in galena crystals from Mississippi Valley-type deposits. Econ Geol 81:1307-1321 Deloule E, Albarede F, Sheppard SMF (1991a) Hydrogen isotope heterogeneities in the mantle from ion probe analysis of amphiboles from ultramafic rocks. Earth Planet Sci Letters 105:543-553 Deloule E, France-Lanord C, Albarede F (1991b) D/H analysis of minerals by ion probe. In Taylor HP, O'Neil JR, Kaplan IR (eds) Stable Isotope Geochemistry. Geochem Soc Spec Publ 3:53-62 Dennis PF (1984a) Oxygen self-diffusion in quartz under hydrothermal conditions. J Geophys Res 89:4047-4057 Dennis PF (1986) Oxygen self diffusion in quartz. Prog Exp Petrol, NERC Publ D 25:260-265 Des Marais DJ, Moore JG (1984) Carbon and its isotopes in mid-oceanic basaltic glasses. Earth Planet Sci Letters 69:43-57 Desmons J, O'Neil JR (1978) oxygen and hydrogen isotope compositions of eclogites and associated rocks from the eastern Sesia Zone (Western Alps, Italy). Contrib Mineral Petrol 67:79-85 Dodson MH (1973) Closure temperature in cooling geochronological and petrologic systems. Contrib Mineral Petrol 40:259-274 Dodson MH (1979) Theory of cooling ages. In Jager E, Hunziker JC (eds) Lectures in Isotope Geology. Springer-Verlag, p 194-202 Doremus RH (1998) Diffusion of water and oxygen in quartz: Reaction-diffusion model. Earth Planet Sci Letters 163:43-51 Dunn SR, Valley JW (1992) Calcite-graphite isotope thermometry: A test for polymetamorphism in marble, Tudor gabbro aureole, Ontario, Canada. J Metamor Geol 10:487-501 Dunn SR, Valley JW (1996) Polymetamorphic fluid-rock interaction of the Tudor Gabbro and adjacent marble, Ontario. Am J Sci 296:244-295 Edwards KJ, Valley JW (1998) Oxygen isotope diffusion and zoning in diopside: The importance of water fugacity during cooling. Geochim Cosmochim Acta 62:2265-2277 Eichmann R, Schidlowski M (1975) Isotopic fractionation between coexisting organic carbon-carbonate pairs in Pre-Cambrian sediments. Geochim Cosmochim Acta 39:585-595 Eiler JM, Valley JW (1994) Preservation of premetamorphic oxygen isotope ratios in granitic orthogneiss from the Adirondack Mountains, New York, USA. Geochim Cosmochim Acta 58:5525-5535 Eiler JM, Baumgartner LP, Valley JW (1992) Intercrystalline stable isotope diffusion: A fast grain boundary model. Contrib Mineral Petrol 112:543-557 Eiler JM, Baumgartner LP, Valley JW (1993) A new look at stable isotope thermometry. Geochim Cosmochim Acta 57:2571-2583 Eiler JM, Baumgartner LP, Valley JW (1994) Fast grain boundary: A Fortran-77 program for calculating the effects of retrograde interdiffusion of stable isotopes. Computers Geosciences 20:1415-1434 Eiler JM, Graham CW, Valley JW (1997a) SIMS analysis of oxygen isotopes: Matrix effects in complex minerals and glasses. Chem Geol 138:221-244 Eiler JM, Valley JW, Graham CM (1997b) Standardization of SIMS analysis of O and C isotope ratios in carbonate from ALH84001. 28th Lunar Planet Sci Conf, p 327-328 Eiler JM, Valley JW, Graham CM, Baumgartner LP (1995a) The oxygen isotope anatomy of a slowly cooled metamorphic rock. Am Mineral 80:757-764 Eiler JM, Valley JW, Graham CM, Baumgartner LP (1995b) Ion microprobe evidence for the mechanisms of stable isotope retrogression in high-grade metamorphic rocks. Contrib Mineral Petrol 18:365-378 Eldridge CS, Compston W, Williams IS, Both RA, Walshe JL, Ohmoto H (1988) Sulfur isotope variability in sediment-hosted massive sulfide deposits as determined using the ion microprobe SHRIMP: I. An example from the Rammelsberg orebody. Econ Geol 83:443-449 Elphick SC, Graham CM (1988) The effect of hydrogen on oxygen diffusion in quartz: Evidence for fast proton transients? Nature 335:243-245 Elphick SC, Graham CM, Dennis PF (1988) An ion probe study of anhydrous oxygen diffusion in anorthite: A comparison with hydrothermal data and some geological implications. Contrib Mineral Petrol 100:490-495 Elsenheimer DW (1988) Petrologic and Stable Isotopic Characteristics of Graphite and Other Carbonbearing Minerals in Sri Lankan Granulites. MS Thesis, University of Wisconsin, Madison

Stable Isotope Thermometry at High Temperatures

405

Elsenheimer DW (1992) Development and Application of Laser Microprobe Techniques for Oxygen Isotope Analysis of Silicates and, Fluid/Rock Interaction During and After Granulite-Facies Metamorphism, Highland Southwestern Complex, Sri Lanka. PhD Thesis, University of Wisconsin, Madison Elsenheimer DW, Valley JW (1992) In situ oxygen isotope analysis of feldspar and quartz by Nd: YAG laser microprobe. Chem Geol Isotope Geosci Section 101:21-42 Elsenheimer DW, Valley JW (1993) Submillimeter scale zonation of δ18O in quartz and feldspar, Isle of Skye, Scotland. Geochim Cosmochim Acta 57:3669-3676 Esler MB, Griffith DWT, Wilson SR, Steele LP (2000) Precision trace gas analysis by FT-IR spectroscopy. 2. The 13C/12C isotope ratio of CO2. Analyt Chem 72:216-221 Essene EJ (1989) The current status of thermobarometry in metamorphic rocks. In Daly JS, Cliff RA, Yardley BWD (eds) Evolution of Metamorpic Belts, p 1-44 Farquhar J, Chacko T (1994) Exsolution-enhanced oxygen exchange: Implications for oxygen isotope closure temperatures in minerals. Geol 22:751-754 Farquhar J, Rumble D (1998) Comparison of oxygen isotope data obtained by laser fluorination of olivine with KrF excimer laser and CO2 laser. Geochim Cosmochim Acta 62:3141-3149 Farquhar J, Chacko T, Frost BR (1993) Strategies for high-temperature oxygen isotope thermometry: A worked example from the Laramie Anorthosite Complex, Wyoming, USA. Earth Planet Sci Letters 117:407-422 Farquhar J, Chacko T, Ellis DJ (1996) Preservation of oxygen isotope compositions in granulites from Northwestern Canada and Enderby Land, Antarctica: Implications for high-temperature isotopic thermometry. Contrib Mineral Petrol 125:213-224 Farquhar J, Hauri E, Wang J (1999) New insights into carbon fluid chemistry and graphite precipitation: SIMS analysis of granulite facies graphite from Ponmudi, South India. Earth Planet Sci Letters 171:607-621 Farver JR (1989) Oxygen self-diffusion in diopside with application to cooling rate determinations. Earth Planet Sci Letters 92:386-396 Farver JR (1994) Oxygen self-diffusion in calcite: Dependence on temperature and water fugacity. Earth Planet Sci Letters 121:575-587 Farver JR, Giletti BJ (1985) Oxygen diffusion in amphiboles. Geochim Cosmochim Acta 49:1403-1411 Farver JR, Giletti BJ (1989) Oxygen and strontium diffusion kinetics in apatite and potential applications to thermal history determinations. Geochim Cosmochim Acta 53:1621-1631 Farver JR, Yund RA (1990) The effect of hydrogen, oxygen, and water fugacity on oxygen diffusion in alkali feldspar. Geochim Cosmochim Acta 54:2953-2964 Farver JR, Yund RA (1991) Oxygen diffusion in quartz: Dependence on temperature and water fugacity. Chem Geol 90:55-70 Farver JR, Yund RA (1991) Measurement of oxygen grain boundary diffusion in natural, fine-grained, quartz aggregates. Geochim Cosmochim Acta 55:1597-1607 Farver JR, Yund RA (1992) Oxygen diffusion in a fine-grained quartz aggregate with wetted and nonwetted microstructures. J Geophys Res 97:14,017-14,029 Farver JR, Yund RA (1995) Grain boundary diffusion of oxygen, potassium and calcium in natural and hotpressed feldspar aggregates. Contrib Mineral Petrol 118:340-355 Farver JR, Yund RA (1995) Interphase boundary diffusion of oxygen and potassium in K-feldspar/quartz aggregates. Geochim Cosmochim Acta 59:3697-3705 Farver JR, Yund RA (1998) Oxygen grain boundary diffusion in natural and hot-pressed calcite aggregates. Earth Planet Sci Letters 161:189-200 Farver JR, Yund RA (1999) Oxygen bulk diffusion measurements and TEM characterization of a natural ultramylonite: Implications for fluid transport in mica-bearing rocks. J Metamor Geol 17:669-683 Fiebig J, Wiechert U, Rumble D, Hoefs J (1999) High precision, insitu oxygen isotope analysis of quartz using an ArF laser. Geochim Cosmochim Acta 63:687-702 Fitzsimons ICW, Mattey DP (1994) Carbon isotope constraints on volatile mixing and melt transport in granulite-facies migmatites. Earth Planet Sci Letters 134:319-328 Forester RW, Taylor Jr. HP (1976) 18O depleted igneous rocks from the Tertiary complex of the Isle of Mull, Scotland. Earth Planet Sci Letters 32:11-17 Forester RW, Taylor HP (1977) 18O/16O, D/H, and 13C/12C studies of the Tertiary igneous complex of Skye, Scotland. Am J Sci 277:136-177 Fortier SM, Giletti BJ (1989) An empirical model for predicting diffusion coefficients in silicate minerals. Science 245:1481-1484 Fortier SM, Giletti BJ (1991) Volume self-diffusion of oxygen in biotite, muscovite, and phlogopite micas. Geochim Cosmochim Acta 55:1319-1330

406

Valley

Fortier SM, Luttge A, Satir M, Metz P (1994) Oxygen isotope fractionation between fluorphlogopite and calcite: An experimental investigation of temperature dependence and F-/OH- effects. Eur J Mineral 6:53-65 Fourcade S, Javoy M (1973) Rapports 18O/16O dans les roches du vieux socle catazonal d'ln Ouzzal (Sahara Algerien). Contrib Mineral Petrol 42:235-244 Freer R, Dennis PF (1982) Oxygen diffusion studies. I. A preliminary ion microprobe investigation of oxygen diffusion in some rock-forming minerals. Mineral Mag 45:179-192 Friedman I, O'Neil JR (1977) Compilation of Stable Isotope Fractionation Factors of Geochemical Interest. U S Geol Surv Prof Paper 440-KK Galimov EM, Rozen OM, Belomestnykh AV, Zlobin VL, Kromtsov IN (1990) The Nature of Graphite in Anabar-Shield Metamorphic Rocks. Geokhimiya 3:373-384 Gerdes ML, Valley JW (1994) Fluid flow and mass transport at the Valentine wollastonite deposit, Adirondack Mountains, N.Y. J Metamor Geol 12:589-608 Ghent ED, Valley JW (1998) Oxygen isotope study of quartz-Al2SiO5 pairs from the Mica Creek area, British Columbia: Implications for the recovery of peak metamorphic temperatures. J Metamor Geol 16:223-230 Giletti BJ (1986) Diffusion effects on oxygen isotope temperature of slowly cooled igneous and metamorphic rocks. Earth Planet Sci Letters 77:218-228 Giletti BJ, Yund RA (1984) Oxygen diffusion in quartz. J Geophys Res 89:4039-4046 Giletti BJ, Hess KC (1988) Oxygen diffusion in magnetite. Earth Plant Sci Letters 89:115-122 Giletti BJ, Semet MP, Yund RA (1978) Studies in diffusion—III. Oxygen in feldspars: An ion microprobe determination. Geochim Cosmochim Acta 42:45-57 Godfrey JD (1962) The deuterium content of hydrous minerals from the East-Central Sierra Nevada and Yosemite National Park. Geochim Cosmochim Acta 26:1215-1245 Gold T (1987) Power from the Earth. Dent & Sons Ltd, London, Melbourne, 208 p Goldman DS, Albee AL (1977) Correlation of Mg/Fe partitioning between garnet and biotite with 18O/18O partitioning between quartz and magnetite. Am J Sci 277:750-767 Graham CM, Valley JW, Winter BL (1996) Ion microprobe analysis of 18O/16O in authigenic and detrital quartz in St. Peter sandstone, Michigan Basin and Wisconsin Arch, USA: Contrasting diagenetic histories. Geochim Cosmochim Acta 24:5101-5116 Graham CM, Valley JW, Eiler JM, Wada H (1998) Timescales and mechanisms of fluid infiltration in a marble: An ion microprobe study. Contrib Mineral Petrol 132:371-389 Gregory RT, Criss RE (1986) Isotopic exchange in open and closed systems. Rev Mineral 16:91-127 Gregory RT, Taylor HP (1986a) Possible non-equilibrium oxygen isotope effects in mantle nodules, an alternative to the Kyser-O'Neil-Carmichael 18O/16O geothermometer. Contrib Mineral Petrol 93: 114-119 Gregory RT, Taylor HP (1986b) Non-equilibrium, metasomatic 18O/16O effects in upper mantle mineral assemblages. Contrib Mineral Petrol 93:124-135 Gregory RT, Criss RE, Taylor HP (1989) Oxygen isotope exchange kinetics of mineral pairs in closed and open systems: Applications to problems of hydrothermal alteration of igneous rocks and Precambrian iron formations. Chem Geol 75:1-42 Hamza MS, Epstein S (1980) Oxygen isotopic fractionation between oxygen of different sites in hydroxylbearing silicate minerals. Geochim Cosmochim Acta 44:173-182 Harte B, Otter ML (1992) Carbon isotope measurements on diamonds. Chem Geol 101:177-183 Hayes JM, Kaplan IR, Wedeking KW (1983) Precambrian organic geochemistry: Preservation of the record. In Schopf JW (ed) Earth's Earliest Biosphere: Its Origin and Evolution. Princeton University Press, Princeton, NJ, p 93-134 Hervig RL, Williams P, Thomas RM, Schauer SN, Steele IM (1992) Microanalysis of oxygen isotopes in insulators by secondary ion mass spectrometry. Int’l J Mass Spect Ion Proc 120:45-63 Hildreth W, Christiansen RL, O'Neil JR (1984) Catastrophic isotopic modification of rhyolitic magma at times of caldera subsidence, Yellowstone plateau volcanic field. J Geophys Res 89:8339-8369 Hoefs J (1997) Stable Isotope Geochemistry, 4th edn. Springer-Verlag, Berlin, 201 p Hoefs J, Frey M (1976) The isotopic composition of carbonaceous matter in a metamorphic profile from the Swiss Alps. Geochim Cosmochim Acta 40:945-951 Hoefs J, Muller G, Schuster AK (1982) Polymetamorphic relations in iron ores from the iron quadrangle, brazil: The correlation of oxygen isotope variations with deformation history. Contrib Mineral Petrol 79:241-251 Hoefs J, Mueller G, Schuster KA, Walde D (1987) The Fe-Mn ore deposits of Urucum, Brazil; an oxygen isotope study. Chem Geol, Isotope Geosci Section 65:311-319 Hoernes S, Friedrickson H (1978) Oxygen and hydrogen isotope study of the polymetamorphic area of the Northern Otztal-Stubal Alps. Contrib Mineral Petrol 67:305-315

Stable Isotope Thermometry at High Temperatures

407

Hoernes S, Lichtenstein U, van Reenen DD, Mokgatlha K (1995) Whole-rock/mineral O-isotope fractionations as a tool to model fluid-rock interaction in deep seated shear zones of the Southern Marginal Zone of the Limpopo Belt, South Africa. S Afr J Geol 98:488-497 Hoffbauer R, Hoernes S, Fiorentini E (1994) Oxygen isotope thermometry based on a refined increment method and its applications to granulite-grade rocks from Sri Lanka. In Raith M, Hoernes S (eds) Tectonic, Metamorphic and Isotopic Evolution of Deep Crustal Rocks, With Special Emphasis on Sri Lanka, p 199-220 Hoffbauer R, Spiering B (1994) Petrologic phase equilibria and stable isotope fractionations of carbonatesilicate parageneses from granulite-grade rocks of Sri Lanka. Precambrian Res 66:325-349 James HL, Clayton RN (1962) Oxygen isotope fractionation in metamorphosed iron formations of the Lake Superior region and in other iron-rich rocks. In Petrologic Studies: Buddington Volume. Geol Soc Am, p 217-239 Jamtveit B, Hervig RL (1994) Constraints on transport and kinetics in hydrothermal systems from zoned garnet crystals. Science 263:505-508 Jamtveit B, Bucher-Nurminen K, Stijfhoorn DE (1992a) Contact metamorphism of layered shale carbonate sequences in the Oslo Rift: I. Buffering infiltration and mechanisms of mass transport. J Petrol 33: 377-422 Jamtveit B, Grorud HF, Bucher-Nurminen K (1992b) Contact metamorphism of layered shale - carbonate sequences in the Oslo Rift: II. Migration of isotopic and reaction fronts around cooling plutons. Earth Planet Sci Letts 114:131-148 Jaszczak J, Robinson G (1998) Spherical graphite from Gooderham, Ontario. 17th Int’l Mineral Assoc Meeting, Toronto, Aug 1998 (abstr) Javoy M (1977) Stable isotopes and geothermometry. J Geol Soc London 133:609-636 Jenkin GRT, Linklater C, Fallick AE (1991) Modeling of mineral δ18O values in an igneous aureole: Closed-system model predicts apparent open-system δ18O values. Geology 19:1185-1188 Jenkin GRT, Fallick AE, Farrow CM, Bowes GE (1991) Cool: A Fortran-77 computer program for modeling stable isotopes in cooling closed systems. Computers and Geosciences 17:391-412 Johnson EA, Rossman GR, Valley JW (2000) Correlation between OH content and oxygen isotope diffusion rate in diopsides from the Adirondack Mountains, New York. Geol Soc Am Abstr Progr 32:A-114 Jones AM, Iacumin P, Young ED (1999) High resolution δ18O analysis of tooth enamel phosphate by isotope ratio monitoring gas chromatography and UV laser fluorination. Chem Geol 153:241-248 Karlsson HR, Clayton RN (1990) Oxygen and hydrogen isotope geochemistry of zeolites. Geochim Cosmochim Acta 54:1369-1386 Kawabe I (1978) Calculation of oxygen isotope fractionation in quartz-water system with special reference to the low temperature fractionation. Geochim Cosmochim Acta 42:613-621 Kelley SP, Fallick AE (1990) High precision spatially resolved analysis of δ34S in sulphides using a laser extraction technique. Geochim Cosmochim Acta 54:883-888 Kemp WM (1985) A Stable Isotope and Fluid Inclusion Study of the Contact Al(Fe)-Ca-Mg-Si Skarns in the Alta Stock Aureole, Alta, Utah. MS Thesis, University of Utah, Provo Kerrich R, Beckinsale RD, Durham JJ (1977) The transition between deformation regimes dominated by intercrystalline diffusion and intracrystalline creep evaluated by oxygen isotope thermometry. Tectonophysics 38:241-257 Kerstel ERT, van Trigt R, Dam N, Reuss J, Meijer HAJ (1999) Simultaneous determination of the 2H/1H, 17 16 O/ O, and 18O/16O isotope abundance ratios in water by means of laser spectrometry. Analyt Chem 71:5297-5303 King EM, Valley JW, Davis DW, Edwards G (1998) Oxygen isotope ratios of Archean plutonic zircons from granite-greenstone belts of the Superior Province: Indicator of magmatic source. Precambrian Res 92:47-67 King EM, Valley JW, Davis DW, Kowallis BJ (2001) Empirical determination of oxygen isotope fractionation factors for titanite with respect to zircon and quartz. Geochim Cosmochim Acta, in press Kitchen NE, Valley JW (1995) Carbon isotope thermometry in marbles of the Adirondack Mountains, New York. J Metamor Geol 13:577-594 Knauth LP, Lowe DR (1978) Oxygen isotope geochemistry of cherts from the Onverwacht group (3.4 billion years), Transvaal, South Africa, with implications for secular variations in the isotopic composition of cherts. Earth Planet Sci Letters 41:209-222 Kohn MJ (1999) Why most “dry” rocks should cool “wet.” Am Mineral 84:570-580 Kohn MJ, Valley JW (1994) Oxygen isotope constraints on metamorphic fluid flow, Townshend Dam, Vermont, USA. Geochim Cosmochim Acta 58:5551-5566 Kohn MJ, Valley JW (1998) Obtaining equilibrium oxygen isotope fractionations from rocks: Theory and examples. Contrib Mineral Petrol 132:209-224

408

Valley

Kohn MJ, Schoeninger MJ, Valley JW (1996) Herbivore tooth oxygen isotope compositions: Effects of diet and physiology. Geochim Cosmochim Acta 60:3889-3896 Kohn MJ, Spear FS, Valley JW (1997) Dehydration-melting and fluid recycling during metamorphism: Rangeley Formation, New Hampshire, USA. J Petrol 9:1255-1277 Kohn MJ, Valley JW, Elsenheimer D, Spicuzza M (1993) Oxygen isotope zoning in garnet and staurolite: Evidence for closed system mineral growth during regional metamorphism. Am Mineral 78:988-1001 Kreulen R, van Beek PCJM (1983) The calcite-graphite isotope thermometer; data on graphite bearing marbles from Naxos, Greece. Geochim Cosmochim Acta 47:1527-1530 Kronenberg AK, Yund RA, Giletti BJ (1984) Carbon and oxygen diffusion in calcite: Effects of Mn content and PH2O. Phys Chem Minerals 11:101-112 Krylov DP, Mineev, S.D. (1994) The concept of model-temperature in oxygen isotope geochemistry: An example of a single outcrop from the Rayner Complex (Enderby Land, East Antarctica). Geochim Cosmochim Acta 58:4465-4473 Kyser TK (1986) Stable isotope variations in the mantle. Rev Mineral 16:141-164 Labotka TC, Cole DR, Riciputi LR (2000) Diffusion of C and O in calcite at 100 MPa. Am Mineral 85:488-494 Larson TE, Sharp ZD (2000) Isotopic disequilibrium in the classic triple point localities of New Mexico. Geol Soc Am Annual Meeting, p A-297 Lasaga AC (1998) Kinetic Theory in the Earth Sciences. Princeton University Press, Princeton, New Jersey, 811 p Layne GD, Longstaffe FJ, Spooner ETC (1991) The JC tin skarn deposit, southern Yukon Territory: II. A carbon, oxygen, hydrogen, and sulfur stable isotope study. Econ Geol 86:48-65 Layne GD, Spooner ETC (1991) The JC tin skarn deposit, southern Yukon Territory: I. Geology, paragenesis, and fluid inclusion microthermometry. Econ Geol 86:29-47 Li H, Schwarcz HP, Shaw DM (1991) Deep crustal oxygen isotope variations: The Wawa-Kapuskasing crustal transect, Ontario. Contrib Mineral Petrol 107:448-458 Liogys VA, Jenkins DM (2000) Hornblende geothermometry of amphibolite layers of the Popple Hill gneiss, north-west Adirondack Lowlands, New York, USA. J Metamor Geol 18:513-530 Loucks RR (1996) Restoration of the elemental and stable-isotopic compositions of diffusionally altered minerals in slowly cooled rocks. Contrib Mineral Petrol 124:346-358 Luque FJ, Pasteris JD, Wopenka B, Rodas M, Barrenechea JF (1998) Natural fluid-deposited graphite: Mineralogical characteristics and mechanisms of formation. Am J Sci 298:471-498 Luz B, Barkan E, Bender ML, Thiemens MH, Boering KA (1999) Triple-isotope composition of atmospheric oxygen as a tracer of biosphere productivity. Nature 400:547-550 Manning JR (1974) Diffusion kinetics and mechanisms in simple crystals. In Hofmann AW et al. (eds) Geochemical Transport and Kinetics. Carnegie Inst Washington, Washington, DC, p 3-13 Matthews A (1994) Oxygen isotope geothermometers for metamorphic rocks. J Metamor Geol 12:211-219 Matthews A, Schliestedt M (1984) Evolution of the blueschist and greenschist facies rocks of Sifnos, Cyclades, Greece. Contrib Mineral Petrol 88:150-163 Matthews A, Beckinsale RD, Durham JJ (1979) Oxygen isotope fractionation between rutile and water and geothermometry of metamorphic eclogites. Mineral Mag 43:406-413 McConnell JDC (1995) The role of water in oxygen isotope exchange in quartz. Earth Planet Sci Letters 136:97-107 McCrea JM (1950) On the isotopic chemistry of carbonates and a paleotemperature scale. J Chem Phys 18:849-857 McKibben MA, Riciputi LR (1998) Sulfur Isotopes by Ion Microprobe. In McKibben MA, Shanks III WC, Ridley WI (eds) Applications of Microanalytical Techniques to Understanding Mineralizing Processes. Soc Econ Geol Rev Econ Geol 7:121-140 McLelland J, Chiarenzelli J, Perham A (1992) Age, field, and petrological relationships of the Hyde School gneiss, Adirondack Lowlands, New York: Criteria for an intrusive igneous origin. J Geol 100:69-90 Merritt DH, Hayes JM (1994) Factors controlling precision and accuracy in isotope-ratio-monitoring mass spectrometry. Analyt Chem 66:2336-2347 Moecher DP, Sharp ZD (1999) Comparison of conventional and garnet-aluminosilicate-quartz O isotope thermometry: Insights for mineral equilibrium in metamorphic rocks. Am Mineral 84:1287-1303 Moecher DP, Essene EJ, Valley JW (1992) Stable isotopic and petrological constraints on scapolitization of the Whitestone meta-anorthosite, Grenville, Province, Ontario. J Metamor Geol 10:745-762 Moecher DP, Valley JW, Essene EJ (1994) Extraction and carbon isotope analysis of CO2 from scapolite in deep crustal granulites and xenoliths. Geochim Cosmochim Acta 58:959-967 Mojzsis SJ, Arrhenius G, McKeegan KD, Harrison TM, Nutman AP, Friend RL (1996) Evidence for life on Earth before 3800 million years ago. Nature 384:55-59

Stable Isotope Thermometry at High Temperatures

409

Moore DK, Cherniak DJ, Watson EB (1998) Oxygen diffusion in rutile from 750 to 1000 C and 0.1 to 1000 MPa. Am Mineral 83:700-711 Morikiyo T (1984) Carbon isotopic study on coexisting calcite and graphite in the Ryoke metamorphic rocks, northern Kiso district, central Japan. Contrib Mineral Petrol 87:251-259 Morishita Y, Giletti BJ, Farver JR (1996) Volume self-diffusion of oxygen in titanite. Geochem J 30:71-79 Morrison J, Valley JW (1991) Retrograde fluids in granulites: Stable isotope evidence of fluid migration. J Geol 99:559-570 Morrison J, Barth AP (1993) Empirical tests of carbon isotope thermometry in granulites from southern California. J Metamor Geol 11:789-800 Morrison J, Anderson JL (1998) Footwall refrigeration along a detachment fault: Implications for the thermal evolution of core complexes. Science 279:63-66 Muller G, Schuster A, Hoefs J (1986a) The metamorphic grade of banded iron-formations: Oxygen isotope and petrological constraints. Fortschr Mineral 64:163-185 Muller G, Hans-Joachim L, Hoefs J (1986b) Sauerstoff- und Kohlenstoff-Isotopenuntersuchungen an Mineralen aus gebanderten Eisenerzen und metamorphen Gesteinen nordostlich des Eisernen Vierecks in Brasilien. Geologische Jahrb D 79:21-40 Murphey BF, Nier AO (1941) Variations in the relative abundance of the carbon isotopes. Phys Rev 59:771-772 Nagy KL, Giletti BJ (1986) Grain boundary diffusion of oxygen in a macroperthitic feldspar. Geochim Cosmochim Acta 50:1151-1158 Nier AO (1947) A mass spectrometer for isotope and gas analysis. Rev Sci Instr 18:398 O'Hara KD, Sharp ZD, Moecher DP, Jenkin GRT (1997) The effect of deformation on oxygen isotope exchange in quartz and feldspar and the significance of isotopic temperatures in mylonites. J Geol 103:193-204 O'Neil JR (1986) Theoretical and experimental aspects of isotopic fractionation. Rev Mineral 16:1-41 O'Neil JR, Clayton RN (1964) Oxygen isotope geothermometry. In Craig H, Miller SL, Wasserburg GJ (eds) Isotope and Cosmic Chemistry, p 148-157 O'Neil JR, Pickthorn WJ (1988) Single mineral oxygen isotope thermometry. Chem Geol 71:369 Ohmoto H (1986) Stable isotope geochemistry of ore deposits. Rev Mineral 16:491-556 Pandey UK, Chabria T, Krishnamurthy P, Viswanathan R, Kumar B (2000) Carbon isotope and X-ray diffraction studies on calcite-graphite system in calc-granulites around Usilampatti Area, Madurai, Tamil Nadu. J Geol Soc India 55:37-46 Patterson WP, Smith GR, Lohmann KC (1993) Continental Paleothermometry and Seasonality Using the Isotopic Composition of Aragonitic Otoliths of Freshwater Fishes. In Swart PK, Lohmann KC, McKenzie J, Savin S (eds) Climate Change in Continental Isotopic Records. Geophysical Monogr, p 191-202 Peck WH (2000) Oxygen isotope studies of Grenville Metamorphism and Magmatism. PhD Thesis, University of Wisconsin, Madison Peck WH, Valley JW, Graham CM (2001) Slow oxygen diffusion in igneous zircons from metamorphic rocks (in review) Perry EC, Bonnichsen B (1966) Quartz and magnetite oxygen-18–oxygen-16 fractionation in metamorphosed Biwabik Iron Formation. Science 153:528-529 Perry EC, Ahmad SN (1977) Carbon isotope composition of graphite and carbonate minerals from 3.8-AE metamorphosed sediments, Isukasia, Greenland. Earth Planet Sci Letters 36:280-284 Perry EC, Ahmad SN (1981) Oxygen and carbon isotope geochemistry of the Krivoy Rog iron formation, Ukranian SSR. Lithos 14:83-92 Perry EC, Tan FC, Morey GB (1973) Geology and Stable Isotope Geochemistry of the Biwabik Iron Formation, Northern Minnesota. Econ Geol 68:1110-1123 Perry EC, Ahmad SN, Swulius TM (1978) The oxygen isotope composition of 3,800 M.Y. old metamorphosed chert and iron formation from Isukasia, West Greenland. J Geol 86:223-239 Pineau F, Latouche L, Javoy M (1976) L'origine du graphite et les fractionnements isotopiques du carbone dans les marbres metamorphiques des Gour Oumelalen (Ahaggar, Algerie), des Adirondacks (New Jersey, U.S.A.), et du Damara (Namibie, Sud-Ouest africain). Bull Soc Geol France 7 t. XVIII: 1713-1723 Polyakov VB, Kharlashina NN (1994) Effect of pressure on equilibrium isotopic fractionation. Geochim Cosmochim Acta 58:4739-4750 Polyakov VB, Kharlashina NN (1995) The use of heat capacity data to calculate isotope fractionation between graphite, diamond, and carbon dioxide: A new approach. Geochim Cosmochim Acta 59:25612572 Putlitz B, Valley JW, Matthews A (2001) Oxygen isotope thermometry of quartz-Al2SiO5 veins in high grade metamorphic rocks on Naxos. Contrib Mineral Petrol (in press)

410

Valley

Rafter TA (1957) Sulphur isotopic variations in nature: Part I. The preparation of sulphur dioxide for mass spectrometer examination. New Zealand J Sci Techn B38:849 Rathmell MA, Streepey M, M., Essene EJ, van der Pluijm BA (1999) Comparison of garnet-biotite, calcitegraphite, and calcite-dolomite thermometry in the Grenville Orogen; Ontario, Canada. Contrib Mineral Petrol 134:217-231 Rees CE (1978) Sulphur isotope measurements using SO2 and SF6. Geochim Cosmochim Acta 42:383-389 Riciputi LR (1996) A comparison of extreme energy filtering and high mass resolution techniques for measurement of 34S/32S ratios by ion microprobe. Rapid Comm Mass Spectrom 10:282-286 Robinson BW, Kusakabe M (1975) Quantitative preparation of sulphur dioxide for 34S/32S analyses from sulphides by combustion with cuprous oxide. Analyt Chem 47:1179 Rosenbaum J, Sheppard SMF (1986) An isotopic study of siderites, dolomites and ankerites at high temperatures. Geochim Cosmochim Acta 50:1147-1150 Rosenbaum JM, Mattey D (1995) Equilibrium garnet-calcite oxygen isotope fractionation. Geochim Cosmochim Acta 59:2839-2842 Rumble D (1978) Mineralogy, petrology, and oxygen isotopic geochemistry of the Clough Formation, Black Mountain, Western New Hampshire, U.S.A. J Petrol 19:317-340 Rumble D, Sharp ZD (1998) Laser microanalysis of silicates for 18O/17O/16O and of carbonates for 18O/16O and 13C/12C ratios. In McKibben MA, Shanks III WC, Ridley WI (eds) Soc Econ Geol Rev Econ Geol 7:99-119 Rumble D, Yui T-F (1998) The Qinglongshan oxygen and hydrogen isotope anomaly near Donghai in Jiangsu Province, China. Geochim Cosmochim Acta 62:3307-3321 Rye RO, Bethke PM, Wasserman MD (1992) The stable isotope geochemistry of acid sulfate alteration. Econ Geol 87:225-262 Ryerson FJ, McKeegan KD (1994) Determination of oxygen self-diffusion in akermanite, anorthite, diopside, and spinel: Implications for oxygen isotopic anomalies and the thermal histories of Ca-Alrich inclusions. Geochim Cosmochim Acta 58:3713-3734 Santosh M, Wada H (1993) A carbon isotope study of graphites from the Kerala Khondalite Belt, Southern India: Evidence for CO2 infiltration in granulites. J Geol 101:643-651 Satish-Kumar M (2000) Ultrahigh-temperature metamorphism in Madurai granulites, Southern India: Evidence from carbon isotope thermometry. J Geol 108:479-486 Satish-Kumar M, Wada H (2000) Carbon isotopic equilibrium between calcite and graphite in Skallen Marbles, East Antarctica: Evidence for the preservation of peak metamorphic temperatures. Chem Geol 166:173-182 Satish-Kumar M, Santosh M, Wada H (1997) Carbon isotope thermometry in marbles of Ambasamudram, Kerala Khondalite Belt, southern India. J Geol Soc India 49:523-532 Satish-Kumar M, Yoshida M, Wada H, Niitsuma N, Santosh M (1998) Fluid flow along microfractures in calcite from a marble from East Antarctica: Evidence from gigantic oxygen isotopic zonation. Geology 26:251-254 Savin SM, Lee M (1988) Isotopic studies of phyllosilicates. Rev Mineral 19:189-223 Scheele N, Hoefs J (1992) Carbon isotope fractionation between calcite, graphite and CO2: An experimental study. Contrib Mineral Petrol 112:35-45 Schidlowski M (2001) Carbon isotopes as biogeochemical recorders of life over 3.8 Ga of Earth history: Evolution of a concept. Precambrian Res 106:117-134 Schidlowski M, Appel PWU, Eichmann R, Junge CE (1979) Carbon isotope geochemistry of the 3.7 × 109 yr. old Isua sediments, West Greenland: Implications for the Archaean carbon and oxygen cycles. Geochim Cosmochim Acta 43:189-199 Schrauder M, Beran A, Hoernes S, Richter W (1993) Constraints on the Origin and the Genesis of Graphite-Bearing Rocks from the Variegated Sequence of the Bohemian Massif (Austria). Mineral Petrol 49:175-188 Schulze DJ, Valley JW, Viljoen KS, Stiefenhofer J, Spicuzza M (1997) Carbon isotope composition of graphite in mantle eclogites. J Geol 105:379-386 Schwarcz HP (1966) Oxygen isotope fractionation between host and exsolved phases in perthite. Geol Soc Am Bull 77:879-882 Schwarcz HP, Clayton RN, Mayeda T (1970) Oxygen isotopic studies of calcareous and pelitic metamorphic rocks, New England. Geol Soc Am Bull 81:2299-2316 Shanks WC, Crowe DE, Johnson C (1998) Sulfur isotope analyses using the laser microprobe. In McKibben RA, Shanks WC, Ridley WI (eds) Soc Econ Geol Rev Econ Geol 7:141-153 Sharma T, Clayton RN (1965) Measurement of O18/O16 ratios of total oxygen of carbonates. Geochim Cosmochim Acta 29:1347-1353 Sharp ZD (1990) A laser-based microanalytical method for the in situ determination of oxygen isotope ratios of silicates and oxides. Geochim Cosmochim Acta 54:1353-1357

Stable Isotope Thermometry at High Temperatures

411

Sharp ZD (1991) Determination of oxygen diffusion rates in magnetite from natural isotopic variations. Geology 19:653-656 Sharp ZD (1992) In situ laser microprobe techniques for stable isotope analysis. Chem Geol 101:3-19 Sharp ZD (1995) Oxygen isotope geochemistry of the Al2SiO5 polymorphs. Am J Sci 295:1058-1076 Sharp ZD, Essene EJ (1991) Metamorphic conditions of an Archean core complex in the northern Wind River Range, Wyoming. J Petrol 32:241-273 Sharp ZD, Jenkin GRT (1994) An empirical estimate of the diffusion rate of oxygen in diopside. J Metamor Geol 12:89-97 Sharp ZD, Cerling TE (1996) A laser GC-IRMS techniques for in situ stable isotope analyses of carbonates and phosphates. Geochim Cosmochim Acta 60:2909-2916 Sharp ZD, Giletti, BJ, Yoder HS (1991) Oxygen diffusion rates in quartz exchanged with CO2. Earth Planet Sci Letters 107:339-348. Sharp ZD, Essene EJ, Smyth JR (1993) Ultra-high temperatures from oxygen isotope thermometry of a coesite-sanidine grospydite. Contrib Mineral Petrol 112:358-370 Sharp ZD, Essene EJ, Hunziker JC (1993) Stable isotope geochemistry and phase equilibria of coesitebearing whiteschists, Dora Maira Massif, western Alps. Contrib Mineral Petrol 114:1-12 Sharp ZD, Frey M, Levi KJT (1995) Stable isotope variations (H, C, O) in a prograde metamorphic Triassic red bed formation, Central Swiss Alps. Schweiz mineral petrogr Mitt 75:147-161 Sharp ZD, O'Neil JR, Essene EJ (1988) Oxygen isotope variations in granulite-grade iron formations: Constraints on oxygen diffusion and retrograde isotopic exchange. Contrib Mineral Petrol 98:490-501 Shelton KL (1983) Composition and origin of ore-forming fluids in a carbonate-hosted porphyry copper and skarn deposit: A fluid inclusion and stable isotope study of Mines Gaspe, Quebec. Econ Geol 78:387-421 Shelton KL, Rye DM (1982) Sulfur isotopic compositions of ores from Mines Gaspe, Quebec: An example of sulfate-sulfide isotopic disequilibria in ore-forming fluids with applications to other porphyry-type deposits. Econ Geol 77:1688-1709 Shemesh A, Kolodny Y, Luz B (1983) Oxygen isotope variations in phosphate of biogenic apatites, II: Phosphorite rocks. Earth Planet Sci Letters 64:405-416 Shemesh A, Kolodny Y, Luz B (1988) Isotope geochemistry of oxygen and carbon in phosphate and carbonate of phosphorite francolite. Geochim Cosmochim Acta 52:2565-2572 Sheppard SMF, Gilg HA (1996) Stable isotope geochemistry of clay minerals. Clay Minerals 31:1-24 Shieh Y-S, Schwarcz HP (1974) Oxygen isotope studies of granite and migmatite, Grenville province of Ontario, Canada. Geochim Cosmochim Acta 38:21-45 Shieh Y-S, Schwarcz HP, Shaw DM (1976) An Oxygen Isotope Study of the Loon Lake Pluton and the Apsley Gneiss, Ontario. Contrib Mineral Petrol 54:1-16 Shimizu M, IIyama JT (1982) Zinc-lead skarn deposits of the Nakatatsu mine, central Japan. Econ Geol 77:1000-1012 Sitzman SD, Banfield JF, Valley JW (2000) Microstructural characterization of metamorphic magnetite crystals with implications for oxygen isotope distribution. Am Mineral 85:14-21 Smalley PC, Stijfhoorn DE, Raheim JH, Dickson JAD (1989) The laser microprobe and its application to the study of C and O isotopes in calcite and aragonite. Sedimentary Geol 65:1-11 Spear FS, Florence FP (1992) Thermobarometry in granulites: Pitfalls and new approaches. Precambrian Res 55:209-241 Stern MJ, Spindel W, Monse EU (1968) Temperature dependence of isotope effects. J Chem Phys 48:29082919 Stoffregen RE, Rye RO, Wasserman MD (1994) Experimental studies of alunite: I, 18O-16O and D-H fractionation factors between alunite and water at 250-450°C. Geochim Cosmochim Acta 58:903-916 Taylor BE, O'Neil JR (1977) Stable isotope studies of metasomatic Ca-Fe-Al-Si skarns and associated metamorphic and igneous rocks, Osgood Mountains, Nevada. Contrib Mineral Petrol 63:1-49 Tennie A, Hoffbauer R, Hoernes S (1998) The oxygen isotope fractionation behavior of kyanite in experiment and in nature. Contrib Mineral Petrol 133:346-355 Thrower PA, Mayer RM (1978) Point defects and self-diffusion in graphite. Physica Status Solidi 47:11-37 Turner DR, Bowman JR (1993) Origin and evolution of skarn fluids, Empire zinc skarns, Central Mining District, New Mexico, USA. Appl Geochem 8:9-36 Urey HC (1947) The thermodynamic properties of isotopic substances. J Chem Soc (1947):562-581 Ustinov VI, Grinenko VA (1985) Intrastructural isotope distribution during mineral formation. Geochem Int’l 22:143-149 Valley JW, O'Neil JR (1980) 13C/12C exchange between calcite and graphite: A possible thermometer in Grenville marbles. Geochim Cosmochim Acta 45:411-419 Valley JW, Graham CM (1991) Ion microprobe analysis of oxygen isotope ratios in granulite facies magnetites: Diffusive exchange as a guide to cooling history. Contrib Mineral Petrol 109:38-52

412

Valley

Valley JW, Graham CM (1993) Cryptic grain-scale heterogeneity of oxygen isotope ratios in metamorpic magnetite. Science 259:1729-1733 Valley JW, Graham CM (1996) Ion microprobe analysis of oxygen isotope ratios in quartz from Skye granite: Healed micro-cracks, fluid flow, and hydrothermal exchange. Contrib Mineral Petrol 124:225234 Valley JW, Chiarenzelli J, McLelland JM (1994) Oxygen isotope geochemistry of zircon. Earth Planet Sci Letters 126:187-206 Valley JW, Kinny PD, Schulze MJ (1998b) Zircon megacrysts from kimberlite: Oxygen isotope heterogeneity in mantle melts. Contrib Mineral Petrol 133:1-11 Valley JW, Kitchen N, Kohn MJ, Niendorf CR, Spicuzza MJ (1995) UWG-2, a garnet standard for oxygen isotope ratio: Strategies for high precision and accuracy with laser heating. Geochim Cosmochim Acta 59:5223-5231 Valley JW, Graham CM, Harte B, Eiler JM, Kinny PD (1998a) Ion Microprobe Analysis of Oxygen, Carbon, and Hydrogen Isotope Ratios. In McKibben MA, Shanks III WC, Ridley WI (eds) Applications of Microanalytical Techniques to Understanding Mineralizing Processes. Soc Econ Geol Rev Econ Geol 7:73-98 Valley JW, Komor SC, Baker K, Jeffrey AWA, Kaplan IR, Raheim A (1988) Calcite crack cements in granite from the Siljan Ring, Sweden: Stable isotopic results. In Boden A, Eriksson KG (eds) Deep Drilling in Crystalline Bedrock. Springer-Verlag, New York, p 156-179 van der Pluijm BA, Carlson KA (1989) Extension in the central metasedimentary belt of the Ontario Grenville: Timing and tectonic significance. Geology 17:161-164 van Haren JLM, Ague JJ, Rye DM (1996) Oxygen isotope record of fluid infiltration and mass transfer during regional metamorphism of pelitic schist, Connecticut, USA. Geochim Cosmochim Acta 60:3487-3504 Vannay J-C, Sharp ZD, Grasemann B (1999) Himalayan inverted metamorphism constrained by oxygen isotope thermometry. Contrib Mineral Petrol 137:90-101 Vogel DE, Garlick GD (1970) Oxygen-isotope ratios in metamorphic eclogites. Contrib Mineral Petrol 28:183-191 Vry J, Brown PE, Valley JW, Morrison J (1988) Constraints on granulite genesis from carbon isotope compositions of cordierite and graphite. Nature 332:66-68 Wada H (1977) Isotopic studies of graphite in metamorphosed carbonate rocks of central Japan. Geochem J 11:183-197 Wada H (1978) Carbon isotopic study on graphite and carbonate in the Kamioka Mining District, Gifu Prefecture, Central Japan, in relation to the role of graphite in the pyrometasomatic ore deposition. Mineral Deposita 13:201-220 Wada H (1988) Microscale isotopic zoning in calcite and graphite crystals in marble. Nature 331:61-63 Wada H, Oana S (1975) Carbon and oxygen isotope studies of graphite bearing carbonates in the Kasuga area, Gifu Prefecture, central Japan. Geochem J 9:149-160 Wada H, Suzuki K (1983) Carbon isotopic thermometry calibrated by dolomite-calcite solvus temperatures. Geochim Cosmochim Acta 47:697-706 Wada H, Enami M, Yanagi T (1984) Isotopic studies of marbles in the Sanbagawa metamorphic terrain, central Shikoku, Japan. Geochem J 18:61-73 Waldron K, Lee MR, Parsons I (1994) The microstructures of perthitic alkali feldspars revealed by hydrofluoric acid etching. Contrib Mineral Petrol 116:360-364 Walker FDL (1990) Ion microprobe study of intragrain micropermeability in alkali feldspars. Contrib Mineral Petrol 106:124-128 Watson EB, Cherniak DJ (1997) Oxygen diffusion in zircon. Earth Planet Sci Letters 148:527-544 Weis PL (1980) Graphite skeleton crystals—A newly recognized morphology of crystalline carbon in metasedimentary rocks. Geology 8:296-297 Weis PL, Friedman I, Gleason JP (1981) The origin of epigenetic graphite: Evidence from isotopes. Geochim Cosmochim Acta 45:2325-2332 Wiechert U, Hoefs J (1995) An excimer laser-based micro analytical preparation technique for in situ oxygen isotope analysis of silicate and oxide minerals. Geochim Cosmochim Acta 59:4093-4101 Wright K, Freer R, Catlow CRA (1995) Oxygen diffusion in grossular and some geological implications. Am Mineral 80:1020-1025 Yapp CJ (1990) Oxygen isotopes in iron (III) oxides 2. Possible constraints on the depositional environment of a Precambrian quartz-hematite banded iron formation. Chem Geol 85:337-344 Young ED (1993) On the 18O/16O record of reaction progress in open and closed metamorphic systems. Earth Planet Sci Letters 117:147-167 Young ED, Fogel ML, Rumble D, Hoering TC (1998) Isotope ratio monitoring of O2 for microanalysis of 18 16 O/ O and 17O/16O in geological materials. Geochim Cosmochim Acta 62:3087-3094

Stable Isotope Thermometry at High Temperatures

413

Zhang Y, Stolper EM, Wasserburg GJ (1991) Diffusion of a multi-species component and its role in oxygen and water transport in silicates. Earth Planet Sci Letters 103:228-240 Zheng Y-F (1991) Calculation of oxygen isotope fractionation in metal oxides. Geochim Cosmochim Acta 55:2299-2307 Zheng Y-F (1992) Discussion on the use of δ–Δ diagram in interpreting stable isotope data. Eur J Mineral 4:635-643 Zheng Y-F (1993) Calculation of oxygen isotope fractionation in hydroxyl-bearing silicates. Earth Planet Sci Letters 120:247-263 Zheng Y-F (1993b) Oxygen isotope fractionation in SiO2 and Al2SiO5 polymorphs: Effect of crystal structure. Eur J Mineral 5:651-658 Zheng Y-F (1993c) Calculation of oxygen isotope fractionation in anhydrous silicate minerals. Earth Planet Sci Letters 120:247-263 Zheng Y-F (1995) Oxygen isotope fractionation in magnetites: Structural effect and oxygen inheritance. Chem Geol 121:309-316 Zheng Y-F (1996) Oxygen isotope fractionations involving apatites: Application to paleotemperature determination. Chem Geol 127:177-187 Zheng Y-F, Fu B (1998) Estimation of oxygen diffusivity from anion porosity in minerals. Geochem J 32:71-89 Zheng Y-F, Simon K (1991) Oxygen isotope fractionation in hematite and magnetite: A theoretical calculation and application to geothermometry of metamorphic iron-formations. Eur J Mineral 3: 877-886 Zheng Y-F, Fu Bin YL, Xiao Y, Shuguang L (1998) Oxygen and hydrogen isotope geochemistry of ultrahigh-pressure eclogites from the Dabie Mountains and the Sulu terrane. Earth Planet Sci Letters 155:113-129 Zheng Y-F, Fu B, Yilin X, Yiliang L, Bing G (1999) Hydrogen and oxygen isotope evidence for fluid-rock interactions in the stages of pre- and post-UHP metamorphism in the Dabie Mountains. Lithos 46: 677-693 Zinner E, Ming T, Anders E (1989) Interstellar SiC in the Murchison and Murray meteorites: Isotopic composition of Ne, Xe, Si, C, and N. Geochim Cosmochim Acta 53:3273-3290

This page left blank.

7

Stable Isotope Transport and Contact Metamorphic Fluid Flow Lukas P. Baumgartner Institute of Geosciences Johannes Gutenberg University of Mainz Johann-Joachim-Becher-Weg 21 55099 Mainz - Germany

John W. Valley Department of Geology and Geophysics University of Wisconsin Madison, Wisconsin 53706 INTRODUCTION Stable isotopes are a powerful tool for deciphering the fluid histories of metamorphic terranes. The nature of fluid flow, fluid sources, and fluid fluxes can be delineated in well-constrained studies. Observed isotopic gradients in metamorphic rocks and minerals can thus shed light on many processes involved in mass-transport including diffusion, recrystallization, fluid infiltration, volatilization, metasomatism, and heat flow. Modeling of fluid flow and mineral exchange kinetics offers greatly enhanced understanding of metamorphic processes that can be tested and refined by application of new microanalytical techniques. This review will concentrate on the principles of stable isotope fluid-rock interaction with an emphasis on fluid-rock interaction and fluid flow in contact metamorphism. Earlier reviews discuss some aspects of regional metamorphism and hydrothermal systems (Valley 1986; Kerrich 1987; Nabelek 1991; Young 1995; Ferry and Gerdes 1998; Bowman 1998). Isotopic studies are especially useful for defining the scale of fluid migration. The intensity of interaction between fluids and the minerals in rocks can be assessed. During metamorphism, the scale of isotopic exchange can vary from less than a micrometer to over 10 kilometers. Many fluid-driven processes are characterized by the degree to which fluid flow is concentrated into zones of high permeability. Thus, the definition of two end-member situations is useful. The flow of a pervasive fluid is distributed throughout the pores in a rock. Pervasive flow can be along grain boundaries or fine-scale crack networks and the effect is to homogenize the chemical potential of all components, including stable isotopes, at a macroscopic scale. In contrast, the flow of a channeled fluid is along vein systems, shear zones or other channelways such as rock contacts or more permeable lithologic units. Channeled flow leads to local chemical heterogeneity, allowing some rocks to remain unaffected while others are extensively infiltrated and modified isotopically. If flow is channeled, open and closed systems can occur in close proximity and one-dimensional flow models are not sufficient (e.g. Gerdes et al. 1995b; Baumgartner et al. 1996; Cartwright and Buick 1996; Bolton et al. 1999). Accurate fluid budgets require knowledge of the degree of channelization, fluid pathways, and the fluid flux. The stable isotope composition of a metamorphic rock is controlled by six factors: (1) the composition of the pre-metamorphic protolith; (2) the effects of volatilization; (3) the temperature of exchange; (4) exchange kinetics; (5) fluid composition, and (6) fluid flux. These factors are best evaluated in studies of contact metamorphism because of the 1529-6466/00/0043-0007$05.00

416

Baumgartner & Valley Abbreviations and Symbols α f −r αL βiO 18 O C _ i c D DL DT Δx , Δy , Δz δff , δfr δif , δir δ13Cff , δ13Cfr δ13Cif , δ13 Cir δ18 Off , δ18 Ofr δ18 Oif , δ18 Oir F F1 , F2 18 ΔF O Φ h g Γ 18 J dif O 18 J dispO 18 J inf O K Dj L μr N Pe P Q rk R rf , Rrf R if , Rri ρf ρr ρo T t V_ V vf, vp x, y, z X _ i x

isotope fractionation factor between fluid and rock longitudinal dispersivity stoichiometry of oxygen in phase i concentration of 18 O in phase I normalized concentration diffusion coefficient longitudinal dispersion coefficient transverse dispersion coefficient small increments in x, y, z direction final isotope value of fluid (f) or rock (r) initial isotope value of fluid (f) or rock (r) final 13 C value of fluid (f) or rock (r) initial 13 C value of fluid (f) or rock (r) final 18 O value of fluid (f) or rock (r) initial 18 O value of fluid (f) or rock (r) mole fraction of element of interest remaining in the rock flux in x-direction through surface 1 or 2 (see Fig. 6, below) change in flux of 18 O porosity of rock fresh water equivalent head gravitational acceleration constant tortuosity of pore space diffusive flux of 18 O dispersive flux of 18 O infiltrative flux of 18 O molar equilibrium constant between solid and fluid for isotope j characteristic length scale of system relative fluid viscosity Peclet number pressure fluid production rate per unit volume reaction rate of the mineral/fluid isotope reaction k final isotope ratio of fluid (f) or rock (r) initial isotope ratio of fluid (f) or rock (r) fluid density relative fluid density fluid density at reference pressure temperature time, or dimensionless time volume molar volume Darcy velocity of the fluid phase average linear pore fluid velocity coordinates mole fraction of species I transposed coordinate

Stable Isotope Transport and Contact Metamorphic Fluid Flow

417

good geologic control and the common occurrence of fluids with distinct initial isotopic compositions usually available in these systems. One example is the large initial 18 difference in δ O between intrusive igneous rocks and carbonates which facilitates the study of fluid sources, CO2/H2O ratios, direction of flow, fluid flux, and temperature. The effect of volatilization on stable isotope composition of rocks is assessed first. In the second section of this chapter, the equations for modeling fluid flow are derived and their use reviewed. A short review of stable isotope systematics in contact aureoles is presented in the final part of the chapter. Special attention is given to the Alta aureole, where detailed measurements and theory have been applied by several groups working in fluid/rock interaction. This focus was chosen because the insight gained from contact metamorphism is necessary to attack the more difficult problems related to regional metamorphism (see Valley 1986). “CLOSED SYSTEM” METAMORPHIC VOLATILIZATION Prograde metamorphism of sediments (and to a lesser degree igneous and metamorphic rocks) causes the liberation of “volatile” components by the reaction of lower temperature, volatile rich minerals. If no externally derived fluids infiltrate the rock, volatilization is often referred to as “closed system” even though it is clear that evolved fluids have left the rock. Dehydration is most common, but decarbonation also occurs in carbonate-bearing lithologies (Ferry and Burt 1982) and desulfidation can locally be important (see also Cartwright and Oliver 2000). Volatilization reactions typically have large, positive volume changes. Hence the volume of the produced fluid and the residual solids is greater than that of the initial solids and pore fluids, creating a fluid overpressure sufficient for fluid expulsion (Hanson 1992; Hanson et al. 1995a,b; Ferry 1995; Connolly 1997; Connolly and Podladchikov 2000). Small, transient fluid over-pressures increase permeability allowing buoyant fluids to rise in the crust (Walther and Orville 1982; Walther and Wood 1984; Connolly 1997). In shallow environments, if permeability is sufficiently high to permit convection (see Etheridge et al. 1983; Brace 1984; Gerdes et al. 1998; Bolton et al. 1999; Cook et al. 1997), it is also possible that hydrothermal fluids will migrate laterally or downward. Most evolved fluids are expelled from their rock of origin. The general absence of voids in metamorphosed rocks shows that the amount of retained fluid is almost nothing. In most rocks, fluid inclusions are the sole remnants of metamorphic fluids. The liberation of metamorphic fluids can have a profound effect on the stable isotope composition of the residual rock. Thus, isotopic ratios provide information about the nature and amounts of volatilization that have occurred. The effects of volatilization can be modeled as one of two end-member equilibrium processes: (1) batch volatilization, where all fluid is evolved before any is permitted to escape and (2) Rayleigh volatilization, where each volatile molecule is immediately isolated from its rock of origin due to steady and perhaps slow expulsion. Most natural processes at high temperatures fall between these extremes and the models provide useful limits. At low temperature and in environments with rapidly changing conditions, isotopic disequilibrium can be important (Lasaga and Rye 1993). BATCH VOLATILIZATION If a rock volatilizes in the absence of infiltration and all evolved fluid equilibrates with the rock before being expelled, then the isotopic ratio of the rock will increase or decrease depending on whether the fluid preferentially partitions the light or the heavy isotope. This process is termed “batch” volatilization (Rumble 1982; Valley 1986). Depletion of the rock in the heavy isotope is most common. Assuming equilibrium, the

418

Baumgartner & Valley

magnitude of this effect will vary directly with the amount of volatilization in accord with mass balance:

δrf = δri − (1− F)10001nα f −r

(1)

where F is the mole fraction of the element of interest that remains in the rock after volatilization; αf-r is the fractionation factor (fluid-rock); and δri , δrf are the initial and final isotopic values of the rock in standard permil notation. The amount of 18O depletion relative to 16O caused by batch volatilization of a siliceous dolomite ( δri = 22‰, αf-r = 1.0060 ) is shown by the straight lines in Figure 1. The application of this calculation requires a number of assumptions that will be discussed in more detail later. Most importantly, the large volume increase that accompanies volatilization requires that fluids escape more or less continuously under normal conditions and thus true batch volatilization is unlikely. Nevertheless, the simplicity of Equation (1) makes it very useful. In most instances of volatilization involving the isotopes of oxygen, and for many cases involving C, H and S, the difference between the results of a calculation using Equation (1) and those of a more complex Rayleigh model are minimal compared to other uncertainties involved. In any case, Equation (1) yields a minimum estimate of the isotopic effect due to equilibrium volatilization.

Figure 1. Lowering of δ18O by batch decarbonation (straight line) and Rayleigh decarbonation (curves). F is the mole fraction of oxygen remaining in the rock. Note that for Rayleigh decarbonation, δ18O tends toward –1000‰ if all oxygen is volatilized, but that a calc-silicate limit exists such that F ≥ 0.6 for most metamorphic reactions. There is little difference between the results of the batch and Rayleigh models above F = 0.6 (from Valley 1986).

Stable Isotope Transport and Contact Metamorphic Fluid Flow

419

Rayleigh volatilization The process of Rayleigh volatilization or distillation involves the continuous exchange and removal of infinitely small aliquots of fluid, each before the volatilization of the next (see Rumble 1982; Valley 1986). Rayleigh volatilization may closely approximate the isotopic effects of dehydration and decarbonation (Shieh and Taylor 1969; Rumble 1982; Valley and O’Neil 1984; Valley 1986). The following equations are equivalent and quantify the isotopic change due to Rayleigh volatilization.

Rrf (α f −r −1) i =F Rr

(2)

(

δrf − δri = 1000 F(

α f −r −1)

)

−1

Rrf and Rri are the final and initial isotopic ratios (i.e. and other terms are as defined for Equation (1).

(3) 18

O/16O, etc.) of the rock system

The importance of Rayleigh volatilization is seen from the curves in Figure 1, which solve Equation (3) for αf-r = 1.0060 and δri = 22‰. This calculation is made to model the effect of decarbonation on δ18O in siliceous dolomites, but in general Figure 1 may be applied to any volatilization process involving O, C, H or S (changing the scale on the ordinate axis allows for different values of δi and α. In the case of decarbonation, Figure 1 shows that when F is close to 1.0 the majority of oxygen is still in the rock, only small amounts of CO2 have been liberated, and δ18O of the rock is slightly decreased as the escaping CO2 preferentially partitions 18O over 16O. For values of F ≥ 0.6 the amount of 18 16 O/ O depletion by a Rayleigh process is very similar to that of batch volatilization. In processes where volatilization nears completion, the Rayleigh model becomes important and substantial deviations from the batch model occur. For instance, in Figure 1 at F = 0.05, 95% of the oxygen has been volatilized as CO2 and δ18Orock ≈ 4, which is over 12 permil lower than the corresponding “batch” value. This magnification of the heavy isotope depletion arises because each successive molecule of volatilizing CO2 partitions with an increasingly 18O-depleted rock, compounding the effect. Because of the stoichiometry of most volatilization reactions, values of F below 0.6 are hypothetical for oxygen and are very unlikely in nature (see for example Reactions 4 and 5). This is called the calc-silicate limit (Valley 1986). However, for C, H and S, any value of F from 1.0 to 0.0 may be expected and thus Rayleigh-controlled processes are particularly important for these elements. A rigorous application of Equations (2) and (3) can be complex. At a given moment, αf-r is the stoichiometrically weighted average fractionation between fluid and rock (Rumble 1982). To gain an exact value of αf-r may require knowledge of fractionation factors relative to a mixed-volatile fluid. Further complication arises because αf-r is not constant during reaction- it varies with temperature, the modal abundance of minerals, and the fluid composition. Pressure thus indirectly affects αf-r by controlling temperature of reaction and fluid composition; larger fractionations are expected in low-pressure environments where most metamorphic reactions (with positive P-T slopes) occur at a lower temperature. Furthermore, the exact pre-metamorphic composition of a rock is

420

Baumgartner & Valley

difficult to know and can only be approximated in a regionally metamorphosed terrane. The situation is often better controlled in a contact aureole, where the non-metamorphic protolith is available. Equilibrium values are often accepted as good approximations for α during regional metamorphism, but growth zonation and heterogeneity can be preserved (Kohn et al. 1993; Kohn and Valley 1994). This documents at least partial disequilibrium between the minerals and the fluid. At lower temperatures, minerals are generally not equilibrated during hydrothermal alteration and the approach to equilibrium is uncertain during contact metamorphism. Dehydration Dehydration is the best-known and most common example of metamorphic volatilization. The magnitude of the isotopic effect is controlled by the amounts of water expelled; high-grade rocks contain less water than lower grade equivalents due to progressive dehydration (Spear 1993). In contrast to shales, which are the most water-rich protolith commonly available for metamorphic dehydration (up to 5 wt %), igneous rocks typically contain only 0.5 to 0.8 wt % H2O (Wedepohl 1969).

The effect of dehydration reactions on the value of δ18O in a rock will always be small, less than 1 permil. The magnitude of this change can be calculated with a few simplifying assumptions, all conservatively made so as to maximize the isotope effect. If it is assumed that 5 wt % H2O is driven off, then Equation (1) for batch volatilization can be used with F ≈ 0.9 (H2O is 89 wt % oxygen while rocks are approximately 50%). The temperature effect on αf-r is important due to the crossover in the sign of fractionation. At low temperature (T < 400-500°C) H2O is isotopically lighter than an average rock and dehydration will cause enrichment in 18O/16O in the remaining rock. At T > 500°C, oxygen in H2O is heaver than in most minerals and dehydration causes 18O/16O depletion. Thus, reactions may tend to cancel each other depending on the details of the prograde reaction path. Furthermore, even if the entire 5 wt % H2O is evolved at a low T ( 300°C ) and a large fractionation is chosen ( αf-r = 0.994 ), then Equation (1) yields only δrf − δri = +0.6 permil for oxygen. Given the isotopic heterogeneity of metasedimentary rocks, it is unlikely that such small effects can be recognized or are significant. The effect of dehydration on δD may be much larger because, in contrast to δ18O, values of F can range down to 0.0. The details of this process will depend greatly on the H at 500°C (Suzuoki and value of α, but as an example consider α D/ H2 O −muscovite = 1.018 Epstein 1976). If a rock initially contains 3.0 wt % water and liberates 2.7% then F = 0.1 and from Equation (1) (batch) δrf − δri = -16.2 permil (for the rock). In contrast, as gradual fluid escape is more likely, Rayleigh dehydration (Eqn. 3) yields δrf − δri = -40.6 permil, a substantially larger depletion. Mixed volatile reactions Carbon dioxide is frequently evolved with H2O during metamorphism of siliceous carbonates and marls. Large volumes of rock can be volatilized (up to 40%) with the potential for significant isotopic shifts. In an extreme example of simple decarbonation, the appropriate modal proportions of quartz react with either dolomite or calcite to form a rock that is 100% diopside or wollastonite by reactions such as: (4) CaMg(CO3)2 + 2 SiO2 = CaMgSi2O6 + 2 CO2 dolomite

quartz

diopside

CaCO3 + SiO2 = CaSiO3 + CO2 calcite

quartz wollastonite

(5)

Stable Isotope Transport and Contact Metamorphic Fluid Flow

421

For complete decarbonation, these reactions have volume increases (for rock plus fluid) of approximately +40% at metamorphic P-T conditions showing that fluids must escape during reaction. However, ΔV of reaction for the rock only is ~-35%, potentially creating significant permeability. In spite of the large amounts of reaction indicated by these volume changes, there is a “calc-silicate limit” of F ≥ 0.6 for oxygen (Eqns. 1-3; Fig. 1). This limit exists because the dominant oxygen reservoir of a silicate rock will always be the silicate minerals that remain as the residual metamorphosed rock, even if decarbonation is complete. Many complex reactions or sequences of reactions can be considered in an effort to accurately calculate the isotopic effects of mixed volatile reactions in a metamorphic rock. Sometimes the results are not significantly different from those calculated for Reactions (4) or (5). Volatilization effects always depend on the amounts of fluid evolved by reaction.

Figure 2. Plot of the coupled δ13C vs. δ18O trends that result from batch (straight lines) and Rayleigh (curves) volatilization assuming normal calc-silicate decarbonation (see Fig. 3). Average fractionation factors are appropriate for metamorphic temperatures; α13C(CO2rock) = 1.0022 and α18O(CO2-rock) = 1.006 and 1.012 (larger δ18O shifts correspond to α = 1.012). Values of F(carbon) and the approximate volume reduction of the rock are also shown. The resulting δ18O is little affected by choice of batch versus Rayleigh model, though large differences in estimated δ13C result (from Valley 1986).

COUPLED O-C DEPLETIONS 18

The magnitude of O/16O and 13C/12C depletions is directly linked and can be calculated if reaction stoichiometry is known. Thus, when Equations (1) or (3) are applied to a specific reaction sequence, a coupled O-C trend results, as seen in Figure 2 13 18 where α COC2 −rock = 1.0022, α COO2 −rock = 1.0120 to 1.0060, δ13Cir = 0.0, and δ18Oir = 22.0. 18 The 2 straight lines model batch volatilization (Eqn. 1 and α f −rO = 1.0060, 1.0120). The 2 18 curved lines model Rayleigh volatilization using Equation (3), the same values of α f −rO , and a normal calc-silicate decarbonation trend for F (see discussion for Fig. 3). For δ18O, only small differences (≤1.3 permil) are seen between the Rayleigh and batch calculations, but for δ13C, the limit of the Rayleigh calculation is -1000‰ as F → 0.

422

Baumgartner & Valley

Note that this value is an artifact of the approximation introduced by replacing the isotope ratios in Equation (2) by the delta-values in Equation (3). The ratio Rrf / Rri trends towards 0 (zero). Fcarbon and ΔVreaction for solids are shown on the Rayleigh curves in Figure 2 predicting that large depletions in 13C/12C only occur as decarbonation nears completion and as nearly all carbon is converted to CO2. Thus, the relation of Foxygen vs. Fcarbon controls the shape and magnitude of coupled depletion under Rayleigh conditions. Figure 3 illustrates different relations of Foxygen to Fcarbon, which depend on stoichiometry, and kinetics of reaction. The normal calc-silicate decarbonation trend applies if all minerals in the rock are fully equilibrated during a reaction such as 4 or 5, in which case all carbon in the rock is liberated as CO2 (Fcarbon → 0), while only 40% of the oxygen is released (Foxygen → 0.6). However, if the rock has (for example) 50% excess silicates that are not involved in the reaction, but which still equilibrate isotopically, then Foxygen → 0.8 as Fcarbon → 0.0 along the 50% inert oxygen trend and the amount of 18 16 O/ O depletion will be smaller. Likewise, if 50% of a rock’s carbon does not participate due to stoichiometric excess, a 50% inert carbon trend will be followed and the depletion of 13C/12C will also be less. In practice any trend is possible from ~100% excess oxygen to ~100% excess carbon, but any deviation from normal decarbonation will diminish the total isotopic change in either O or C and the “normal” trend yields the largest isotope variations and is thus quantitatively most important. Depletions in 18O/16O that are larger than those that result from “normal” calcsilicate decarbonation trends may be accomplished by reactions that exceed the calcsilicate limit of Foxygen = 0.6 (Fig. 3). Some rare decarbonation reactions achieve this in the absence of silicates, such as the breakdown of calcite to lime: CaCO3 = CaO + CO2

(6)

which follows a silicate-absent decarbonation trend (Fig. 3). Still larger 18O/16O depletions would be possible if carbonates volatilize without maintaining isotopic equilibrium with coexisting silicates (Lattanzi et al. 1980; Lasaga and Rye 1993).

Figure 3. Values of F(carbon) vs. F(oxygen) along various reaction paths. F is the mole fraction of O or C remaining in the rock after reaction. Most marbles, if equilibrated, will follow a F-F path intermediate between 50% inert oxygen and 50% inert carbon (from Valley 1986).

Stable Isotope Transport and Contact Metamorphic Fluid Flow

423

Coupled O-C isotope depletions are seen in many metamorphic systems involving carbonate rocks. Table 1 and Figures 4 and 5 summarize results for 28 studies of marbles, mostly in contact metamorphic settings. These results will be discussed in more detail later, but one conclusion is general and deserves emphasis. In each of these localities, the O-C trend has a negative slope in Figures 4 and 5, qualitatively similar to the effects of devolatilization. However, in each area, the magnitude of the depletions is too large to be explained by the closed system the devolatilization processes discussed here. Significant fluid infiltration and exchange involving low δ18O, low δ13C fluids is indicated by the stable isotope data (see Valley 1986).

Figure 4. Coupled O-C trends showing decreasing values of δ18O and δ13C with increasing metamorphic grade from localities in Table 1. Trends mimic one another, starting at values for normal marine limestone (δ18O = 20 to 26‰, δ13C = -2 to 4‰) and decreasing towards igneous values. The stippled field, labeled Rayleigh Volatilization, models the effect of volatilization as in Figure 2. Most samples are from contact metamorphic wall rocks or skarns. These trends indicate large amounts of fluid flow and infiltration by magmatic and/or meteoric fluids (Valley 1986).

OPEN SYSTEM FLUID-ROCK INTERACTION: CONTINUUM MECHANICS MODELS OF STABLE ISOTOPE TRANSPORT IN HIGH TEMPERATURE CRUSTAL SYSTEMS

Continuum mechanics models for stable isotope fluid-rock exchange have been developed over the last three decades to accurately describe fluid flow in the Earth’s crust. The geometry of crustal rocks, especially in areas of active metamorphism and deformation requires that fluid flow in the Earth is modeled in three dimensions. Evaluation of such aspects as the importance of fluid flow as driving force for metamorphism (e.g. Fyfe et al. 1978; Ferry 1980, 1986, 2000; Baumgartner and Ferry 1991; Ague 1997), and the influence of metamorphic fluid fluxes on climate (Kerrick and Caldeira 1999) depend on an accurate appraisal of the amount and the timing of fluid percolating through the crust. Significant progress has been made applying continuum

424 

Baumgartner & Valley 

Table 1. Studies demonstrating coupled O-C depletion trends in metamorphosed carbonates. Locality numbers are the same as in Figure 4.

1.

4. 5.

Trenton limestone Mount Royal, Quebec Marysville, Montana Pine Creek, California W-skarn Osgood Mts., Nevada Skye, Scotland

6.

Elkhorn, Montana

7.

Notch Peak Stock, Utah

8.

Weolag W-Mo Deposit, Korea Tauern Area, Austria Birch Creek, California

2. 3.

9. 10. 11. 12. 13. 14. 15. 16. 17.

McArthur R. Pb-Zn Deposits, Australia Providencia Pb-Zn Deposits Mexico Gaspé Cu Deposits, Quebec CanTung W-skarn, N.W. Territories Mottled Zone, Israel Bergell Aureole, Italy Alta Stock, Utah

Width of aureole or traverse 100 m

Pressure/ Depth

Maximum temperature °C

δ18O

δ13C

X(CO2)

range ‰ 14

range ‰ 6

1-3 m roof pendant